Positive

All species for which we had data utilized burn areas (Tables 1 and 2). How and when individuals or species used burn sites depended upon the species (Pearson et al. 1995) and responses were likely related to the resource availability post burn. As most of the studies examined focused on ungulates, we have more nuanced information on burn use for these species. Bighorn sheep (Ovis canadensis) moved back onto burn areas as forage plants began to recover and improved habitat conditions led to population increases and increased use of burned areas for two winters post-fire (Peek et al. 1985). There was also a reduction in sheep density due to increased forage extent, which led to reduced die-offs and reduced effects of lungworm infection in sheep that foraged in burned areas (Peek et al. 1985, Smith et al. 1999). Stone's sheep (Ovis dalli) with access to burned subalpine ranges were in better condition, showed greater horn growth in yearling rams, reduced lungworm infection and higher lamb/ewe ratios compared to those restricted to unburned alpine ranges (Seip & Bunnell 1985).

Burning substantially increased the nutritional quality of winter diets of mountain sheep (Ovis sp.) as well as mule deer and white-tailed deer in both montane grassland and shrubland ecosystems (Anderson 2006). Mule deer and mountain sheep diets were higher in crude protein and digestibility in burned areas (Hobbs & Spowart 1984). Burning also had effects on tongue papillae morphology (length and surface enlargement) and rumen digesta dry weight in white-tailed deer and mule deer within 3 years post-fire (Zimmerman et al. 2006). Increases in body mass and antler measurements of harvested deer after fire were also observed (Johnson et al. 1992). The only study on Equidae responses to fire showed horses preferentially selected areas managed by recent fire due to new growth (Leverkus et al. 2018). Although abundance of hares, mustelids, canids and caribou generally decreased post-fire in response to early seral stages and lack of cover, moose (Alces alces), white-tailed deer, mule deer and elk foraged for the young successional browse in recent burns (Fisher & Wilkinson 2005).

Negative

Studies reporting negative impacts of fire were found for two species only, beaver (Castor canadensis) and polar bear (Ursus maritimus). Two studies on beaver showed that lodge occupancy was significantly lower in burned than in unburned areas. Although beaver populations in some burned areas increased in the years post-fire, original occupancy levels were not recovered over a 12-year post-fire period (Hood et al. 2007). Interestingly, there was an increase in beaver occupancy 2 years after the burn, but it was not sustained, even though food resources were still available to beavers, indicating that occupancy declines may have been driven by more than just fire (Hood & Bayley 2003).

Female polar bears move onto land to den in the northern edges of fire-prone boreal forest (Richardson et al. 2007). Upon returning to denning habitat within 1 year post-fire, two bears did not create dens. Due to unstable frozen peat banks, one pregnant female spent 53 days in burned areas then left to den in unburned forest, and the second female spent 15 days in burned areas and returned to sea ice. There were no polar bears in the burn for 5 years post-fire (Richardson et al. 2007).

Burned areas may act as population or dispersal sinks and provide unsustainable or lower quality habitat for carnivores, such as juvenile or newly dispersed martens and fishers who hunt prey utilizing recent burns, but would not otherwise occur in the younger burns (<10 year old burns; Fisher & Wilkinson 2005).

Mixed

Species may exhibit both positive and negative responses to fire, for example, when feeding habitat is immediately improved post-fire, but reproductive habitat is diminished due to lack of cover. Further, species with high mobility and adaptability may not exhibit observable responses to fire. It has been assumed that fire and burned areas adversely affect caribou because they rely on late several stages (old growth forests) where lichens are abundant and provide winter forage (Luensmann 2007). Although caribou responses to fire can be complex, overall use appears indistinguishable from regular, interannual variation in space use (Table 2) (Silva et al. 2020). Caribou showed overlap in pre- and post-fire seasonal home ranges even during calving. Early seral forage in burned areas attracted caribou similar to other ungulates (Joly et al. 2003), though caribou decreased the use of the burn in winter (Joly et al. 2003, Silva et al. 2020).

Moose use of burned areas illustrates the high variability of burn use even within single species. A comparison of moose habitat use and activity patterns between two large, burned areas in Alaska found that moose used remnant forest more frequently than burned forest (Bangs et al. 1985). When moose did utilize burn areas, they selected burned forest habitat rather than other burned habitat types and remained within 100 m of remnant forest. Moose with low pre-fire use of an area increased after fire use by shifting home ranges to include more of the burn and increasing time spent in the burn, while other moose shifted out of an area post-burn that had high pre-fire utilization and some moose showed no post-fire use despite their close proximity to the burn (Gasaway et al. 1989). Similar to bighorn sheep, there was a population increase of moose following fire, even though some individuals decreased their use of burned habitat (Gasaway et al. 1989).

White-tailed deer commonly exhibit mixed responses to fire, largely owing to their need for concealment cover, particularly when fawning, while still being attracted to the nutritious and palatable forage generated by fire. For example, in one study, while white-tailed deer showed no apparent alarm to fire, they did avoid areas most severely affected by fire (Ivey & Causey 1984). In Texas semi-arid rangeland, white-tailed deer avoided burned areas during dawn and dusk and generally used these areas less (Meek et al. 2008).

Collared peccaries (Pecari tajacu) utilized both burned and unburned chaparral, and unburned desert scrub, but avoided burned desert scrub (O'Brien et al. 2005). The burned and unburned chapparal provided more thermal and concealment cover, particularly when coyotes (Canis latrans) were present. In a longleaf pine forest, raccoons (Procyon lotor) were 62% more likely to use unburned stands over burnt areas (Jones et al. 2004), but in an Oak savanna, raccoon locations, number of visits and amount of time spent were similar in pre-burned and burned areas (Sunquist 1967). The lone study on American badger, whose populations are decreasing (IUCN 2022), failed to detect either a positive or negative response to fire (Holbrook et al. 2016). A single-species study of grey wolf and two studies of mountain lion also did not detect either a positive or negative response to fire (Potter & Kessell 1980, Ballard et al. 2000, Main & Richardson 2002).

Predator–Prey

Fire can influence predator–prey interactions by rapidly modifying the distribution of concealment cover and food resources (Jorge et al. 2020). When coyotes were present, mule deer decreased the use of all burned and unburned desert scrub habitat and increased the use of both burned and unburned chaparral for quality forage, thermal and escape cover (O'Brien et al. 2010). A similar response was observed in collared peccaries, who also increased the use of burned and unburned chaparral versus scrub habitat when coyotes were present (O'Brien et al. 2005).

Wolves utilized a burn area while fire was in progress in northwestern Alaska, likely due to an increased availability of various sized prey as cover burned (Ballard et al. 2000). Wolves continued to use the burned areas more than expected the following summer, but decreased use during winter, apparently the result of shifts in caribou distribution. Fire increased the amount of overlap between wolf and caribou habitat, resulting in both direct (improved forage) and indirect effects (increased predation risk) for caribou (Robinson et al. 2012).

There were no direct effects on the American badger from the altered fire regime caused by invasive cheatgrass (Bromus tectorum), but an indirect effect due to the negative impact of an altered fire regime on the Piute ground squirrel (Urocitellus mollis), a major prey resource (Holbrook et al. 2016). In a study examining mountain lion predation of mule deer, the majority of kill sites occurred in unburned areas, but the number of observed kill sites in burned areas was greater than expected given the extent of burned habitat (Jennings et al. 2016).

Habitat partitioning

In two studies conducted on multiple carnivore species habitat use post-fire, one study showed higher grey fox (Urocyon cinereoargenteus) occupancy in the burn interior, increased striped skunk (Mephitis mephitis) occupancy at burn edges, increased coyote (Canis latrans) abundance with weak associations with the burn interior, and low bobcat (Lynx rufus) detections 3 years post-fire. The four carnivores were, however, found to be resilient to fire (Schuette et al. 2014).

Similarly, ungulates also partitioned the burned landscape such that Stone's sheep most often used younger burns (<3 years) and burned areas during winter, while elk used burns of all ages during all seasons (Sittler et al. 2014, 2015). In a tall-grass prairie ecosystem, elk and reintroduced bison partitioned the landscape such that bison bulls preferred 1-year-old burns (>4 mos.) and mixed groups of young male and female bison preferred recent burns (<4 mos.), while elk preferred unburned areas (Kagima & Fairbanks 2013). Bison foraging in recently burned areas likely forced elk to select unburned forbs and non-native plants (Kagima & Fairbanks 2013). Reduced cover and slow recovery of browse species or high use by elk may have contributed to the avoidance of burned areas by mule deer (Roerick et al. 2019).

Animal responses to fire are also affected by interactions with humans. Fire on the Kenai Peninsula in Alaska showed an initial decrease in moose (Alces alces) populations in the burned area due to increased exposure to anthropogenic hunting rather than reduction in available forage (Spencer & Hakala 1964).

Fire severity and extent

Fire severity impacts how habitats recover from fire and influences how species use burned areas. White-tailed deer utilized high-intensity burn areas more than the surrounding unburned areas but only after 9 years post-burn (Vogl & Beck 1970). Mule deer increased the use of areas with a higher average fire severity, which had a significant positive effect on winter habitat use (Bristow et al. 2020). Conversely, in a study examining the effects of fire size and severity on the distribution and abundance of elk, mule deer and domestic livestock, areas that experienced low-severity fires had the largest number of deer pellets (Wan et al. 2014). A single study on black bear showed the majority of den and bed sites were located within low-severity wildfire burned areas (Bard & Cain 2020). We found very little information on mammal responses to the extent of a burn area, except that bison avoided larger burned patches as smaller patches may be easier for herbivores to keep the plant community in an early successional state of high nutritional value by grazing (Allred et al. 2011).

Naturally occurring versus prescribed fire

Wildfire and prescribed burns are often compared, analysed and interpreted together. However, prescribed burns are of lower intensity, and are orders of magnitude smaller than wildfires (Volkmann et al. 2020). Therefore, small-scale prescribed burns will likely not produce the same effects at different scales, across all habitats, taxa and locations (Pastro et al. 2014). Prescribed fire can be effective for groups with high site fidelity that feed in patches, such as bighorn sheep. Bighorn sheep had a strong positive response by shifting habitat use patterns to include formerly unused areas and moving into areas treated with prescribed fire (Smith et al. 1999). Alpha and beta diversity in vertebrate assemblages (mammals, birds, reptiles and amphibians) was also impacted by burn type (Pastro et al. 2014). Prescribed burns increased alpha diversity, but wildfires did not. However, species assemblages between burned and unburned habitats were less similar after wildfire, indicating higher beta diversity when compared with prescribed burns (Pastro et al. 2014).

Although beaver lodge occupancy increased slightly after one prescribed burn, it decreased dramatically after two burns, and occupancy dropped to zero after three or more burns (Hood & Bayley 2003). Combining multiple perturbations such as drought and extensive or repeated burning appeared to cause more disruption to wetland habitat than beavers were able to tolerate (Hood & Bayley 2003). Black bears selected den and bed sites in unburned and wildfire-burned areas, but no den or bed sites were located in prescribed burned areas (Bard & Cain 2020). Frequent annual prescribed burns across multiple habitat types (e.g. riparian, open, pine, swamp) also decreased the use of prescribed burn sites for bear dens (Stratman & Pelton 2007). Interestingly, Volkmann et al. (2020) found that no carnivore family had more than three studies addressing the effects of prescribed fire, which is surprising considering the prevalence of prescribed fire in many forests where carnivores are common.

The examples provided here illustrate that wildfire and prescribed burns have different effects and are not directly comparable. Prescribed fire may be more effective in some habitats than others, for example, prescribed burning effects on forage were more persistent in a shrub community than grassland habitat (Hobbs & Spowart 1984). While prescribed fire can enhance habitat in patches in the short-term (Pastro et al. 2014), prescribed fire cannot be considered a substitute for wildfire.

Time since fire

While no single fire regime will generate the highest biodiversity, fire effects such as time since fire, fire extent and fire severity help tease apart mechanisms behind mammal responses to fire, such as determining when animals return to areas previously occupied before fire (Van Dyke & Darragh 2006, van Mantgem et al. 2015). Bison and domestic cattle used recently burned areas more than random with site selection driven by time since fire. The amount of time spent in burned areas was greater for native bison than cattle (Allred et al. 2011). There was a strong positive association between bighorn sheep and areas burned up to 15 years post-fire, and a strong negative association between sheep and chaparral >15 years post-fire (Bleich et al. 2008). Caribou were also observed avoiding recent and/or large burned areas, but selected areas that were burned 1–10 and 31–44 years prior, areas closer to small burns (<10,000 ha) and areas within 500 m of the burn perimeter (Anderson & Johnson 2014). Two months post-fire white-tailed deer use of semi-arid rangelands in Texas increased more than expected, followed by a decline in use after 4 months, and use decreased to less than expected after 10 months (Meek et al. 2008). Similarly, in a study of multiple species (white-tailed deer, mountain lion, bobcat, black bear and raccoon) in a Florida broadleaf forest, white-tailed deer populations increased significantly in a recent burn (<6 months), but relative abundance declined at >48 months post-fire (Main & Richardson 2002).

Carnivore habitat use was affected by time since fire. Coyotes selected more recently burned areas (0–1 years) and avoided areas burned ≥4 years prior due to food availability (soft mast or fleshy fruits, and small mammal abundance; Stevenson et al. 2019). Track-counting surveys and counts of trail intersections to examine post-fire habitat selection and abundance of lynx (Lynx canadensis) and snowshoe hares, showed that lynx almost exclusively used areas that had burned ~30 years prior than either mature forest or more recent burn areas (~10 years post-fire; Paragi et al. 1997). Lynx used burned landscapes regularly, more often and earlier post-fire than previously thought, entering burned areas as early as 1 year post-fire, but avoiding areas of recent high-severity fire, and selected fire skips (islands of forest skipped by large fires), high canopy cover and older burns (Vanbianchi et al. 2017).

Mountain lions in California avoided habitats <1 year post-fire and preferred habitat 2–5 years post-fire, with greater positive responses to burned habitat relative to availability within <1–9 years post-fire (Jennings et al. 2016). In Florida, mountain lions increased the use of burned habitats within 1 year after burning again likely due to changing habitat use by prey species (Dees et al. 2001). Black bears have specific habitat requirements for den and bed sites, which can be affected by wildfire. Increased utilization of burned sites was highly contingent on fire severity and vegetation recovery within burn sites over time (Bard & Cain 2020).

Fire suppression and land management

Studies show that direct mortalities from fire for large- and medium-sized mammals are uncommon (Lawrence 1966, Komarek 1969, Ivey & Causey 1984, Singer et al. 1989, Soyumert et al. 2020). However, fire suppression promotes shrub and tree encroachment, which degrades habitat for species that require open areas with high visibility, and fosters extensive even-aged forest stands, which are very vulnerable to blight, insect infestation and large catastrophic crown fires (Smith et al. 1999). Wildfire suppression and conifer encroachment had a significant negative impact on a seral shrubland community, which is important grizzly bear habitat and food source (Zager et al. 1983). In the San Gabriel Mountains in southern California, bighorn sheep experienced a precipitous population decline from 740 sheep in 1980 to 63 in 2002 due to changes in habitat quality (succession of chaparral vegetation and reduced visibility) caused by fire suppression, resulting in increased predation by mountain lions (Bleich et al. 2008, Holl & Bleich 2010). Suppression of natural wildfires likely impacts both vertebrate and invertebrate herbivores that select rapidly growing vegetation (Stein et al. 1992).

The current standard policy of wildfire management has been suppressing 99% of all fires on public lands (Dale 2006). This policy has been in effect for over 100 years even in fire-prone and fire-adapted ecosystems and not only has led to reduced habitat heterogeneity and species diversity but also to an increase in the frequency of catastrophic crown fires that can destroy huge tracks of forest and delay ecosystem recovery.

While important to reduce fuel load in forests that have not burned for decades, prescribed fire management policy should include consideration of timing of fire relative to important biological functions such as the fawn rearing and nesting seasons (Jorge et al. 2020). In Fort Bragg, North Carolina, burns encompass large contiguous and adjacent areas often burned in the same year with average burn block sizes (43 ha) similar to the average core area of female deer during lactation (Lashley et al. 2015). But without cover, female deer become restricted to narrow, linear strips in moist drainages, predisposing females and fawns to predation and unable to take advantage of the nutritious growth provided by the burns (Lashley et al. 2015).

Fire management can have profound impacts to wildlife and their habitat. For example, 38 bighorn sheep were translocated from Oregon to Wyoming, land management agencies recaptured those individuals 1–2 years later (Clapp & Beck 2016), when stress levels may have been high due to relocation and unfamiliarity with the new landscape ecological conditions. In addition, two prescribed burns were implemented in consecutive years. One burn during favourable climatic conditions and the other during drought. These were followed by naturally occurring wildfire, which resulted in a precipitous decrease in transplanted bighorn survival, and greater than 30% mortality of the population (Clapp & Beck 2016). The researchers concluded that although bighorn sheep did expand their range to forage in burned areas, they did not detect active selection of burned areas, suggesting that the expansion may have been an indication that bighorn sheep used burned habitats but only in conjunction with proximity to escape terrain.