Introduction

Fire historically served as a major disturbance agent structuring forests in the western United States (Agee 1993, Covington et al. 1994, Allen et al. 2002, Baker 2015 a). Fire suppression, along with other anthropogenic activities, has altered species composition, forest structure, and fuel complexes in many forest types, leaving these forests vulnerable to high-severity wildfires (Covington and Moore 1994, Allen et al. 2002, Brown et al. 2004). Studies suggest that both size and severity of wildfires has increased in the western US over the past several decades in response to structural changes and changing climate (Littell et al. 2009, Miller et al. 2009, Miller and Safford 2012; but see Odion and Hanson 2006; Williams and Baker 2012; Baker 2015a, b Hanson and Odion 2016), and climatic projections suggest that these trends will continue as future climate becomes warmer and drier (McKenzie et al. 2004, Westerling et al. 2006, Littell et al. 2009, Allen et al. 2010).

This increase in extent and severity of wildfires resulted in a current management emphasis on restoring Western forests to conditions within their historical range of variability (Keane et al. 2009) in hopes of increasing their resiliency to wildfire (Allen et al. 2002, Brown et al. 2004, Stanturf et al. 2014, Hagmann et al. 2017), culminating in large-scale forest restoration projects (Sisk et al. 2005, North et al. 2009, Roccaforte et al. 2010). These efforts are controversial and the subject of debate, however. A species of particular interest in this debate is the spotted owl (Strix occidentalis Xantus de Vesey). All three recognized subspecies of spotted owls (California [S. o. occidentalis], Mexican [S. o. lucida Nelson], and northern [S. o. caurina Merriam]) typically nest in areas of high canopy cover within late seral forests (Forsman et al. 1984; Tempel et al. 2014, 2016; Ganey et al. 2016), although Mexican spotted owls also nest in rocky canyonlands (Willey and van Riper 2007, USDI FWS 2012). Due to stand density and abundance of dead material, most nesting stands have high fuel loads and therefore are vulnerable to high-severity wildfire. This is particularly true in drier forest types, but even mesic nesting stands can be vulnerable when embedded in landscapes dominated by drier forest types that are susceptible to crown fires following decades of fire suppression. Nesting habitat requires many decades to achieve structural conditions conducive to spotted owl nesting, and the loss of nesting habitat to high-severity fire has traditionally been viewed as a primary threat to the spotted owl (USDI FWS 2012). As a result, forest restoration treatments aimed at reducing fuel amounts and continuity have been proposed as a means to reduce risk of loss of spotted owl habitat to wildfire (James 2005, Lee and Irwin 2005, Gaines et al. 2010, USDI FWS 2012, Hagmann et al. 2017). Some authors argue against this approach, however, on the basis that high-severity fire was relatively common in many of these forest types, and large-scale restoration efforts therefore are unnecessary and, in fact, may degrade habitats for threatened and endangered species (Odion and Hanson 2006; Hanson et al. 2009a, b; Williams and Baker 2012; DellaSala et al. 2013; Odion et al. 2014; Baker 2015a, b; Bond 2016; Hutto et al. 2016). Further complicating this debate, the methods used and conclusions reached in some of these studies are not universally accepted among fire and forest ecologists (Fulé et al. 2014, Levine et al. 2017, O’Connor et al. 2017; but see Williams and Baker 2014).

Existing studies indicate that fuels reduction treatments may degrade habitat quality for spotted owls, however (Meiman et al. 2003; Seamans and Gutiérrez 2007; Stephens et al. 2014 a; Tempel et al. 2014, 2015). Consequently, a conflicting viewpoint has emerged, suggesting that fuels-reduction treatments are unnecessary and misguided because they degrade owl habitat, do not reduce the extent of high-severity fire, and may result in greater loss of spotted owl habitat than wildfire alone would cause (Hanson et al. 2009a, b; DellaSala et al. 2013; Odion et al. 2014; Baker 2015 b; Hanson and Odion 2016; but see Spies et al. 2009).

This conflict over whether high-severity wildfires or fuels treatments are potentially more damaging to spotted owls creates a dichotomy of perspectives regarding conserving spotted owls and their habitat. The resulting debate over tradeoffs among spotted owl habitat, wildfire, and restoration treatments has implications for management of vast tracts of forested land supporting numerous resources, as well as for how we integrate conservation of endangered species and their habitat with other resource management objectives (Hanson et al. 2009a, b; Spies et al. 2009; DellaSala et al. 2013; Odion et al. 2014; Baker 2015 b; Hanson and Odion 2016).

Here, we argue that the existing literature is not sufficient to unambiguously quantify the response of spotted owls to high-severity wildfire, and that high-severity fire is pervasive enough within the range of the spotted owl to constitute a potential threat to owl habitat. We also provide evidence that forest restoration and fuels reduction treatments can mitigate fire behavior, but acknowledge that these treatments also can degrade spotted owl habitat. Based on these findings, we argue for cautious implementation of restoration treatments in or near spotted owl habitat, with the goal of identifying treatment types that successfully reduce fire risk while maintaining suitable habitat conditions for spotted owls.

These are complicated issues. Our intent is not to exhaustively review the existing literature on owl response to wildfire (recently reviewed in Bond 2016) or historic fire regimes within the range of the spotted owl, an area of ongoing debate (see above). Rather, we hope to clarify the limits to possible inference and caution against over-extrapolating results from existing studies in this debate. Below, we discuss some key points contributing to uncertainty over the response of spotted owls to high-severity wildfire, and recommend studies that would further our understanding on that response, as well as the response of spotted owls to fuel treatments. We focus our discussion on high-severity wildfire because it generally is perceived as a more pressing concern for spotted owls and their habitat than low-severity or moderate-severity wildfire.

Owl Response Differs Among Studies

Although the recent literature contains numerous studies of the effects of fire on spotted owls, results of these studies do not indicate a consistent response by spotted owls. In cases in which multiple studies examined the same parameter, the results of those studies often showed a mix of responses (Table 1). For example, many studies showed no effects of wildfire on territory occupancy rates, whereas others showed negative effects, and studies that evaluated reproductive rates or use of foraging habitat showed responses ranging from positive to negative (Table 1). Only a single study to date presented long-term data on demographic performance of owls in burned areas, and that study also indicated variability among areas and fires (Rockweit et al. 2017). Thus, available data suggest considerable variation in responses of owls to wildfire.

Table 1 Summary of studies evaluating the response of California (CSO), Mexican (MSO) and northern (NSO) spotted owls to high-severity wildfire. Also shown are parameters evaluated; whether or not salvage logging occurred in burned areas (No, Yes, or Unknown [UN]); number of owls, sites, or territories (T = territories, B = number of burned territories) included; number of years post fire covered by the evaluation; and a simplified response (+ = generally positive, 0 = neutral, − = generally negative). The parameter occupancy was broadly defined here, and often included separate estimates of colonization and extinction rates. Number of territories and fires included, and number of years post fire covered by evaluation were estimated as best we could using information in the papers; that information was sometimes incomplete or unclear.
Full size table

Some of the variation among studies may reflect differences among geographic regions and forest types covered. All three subspecies of spotted owls inhabit forest types ranging from wet to fairly dry. For example, forests inhabited by northern spotted owls range from very mesic in places such as the Olympic Peninsula, Washington, to drier forest types on the eastern slope of the Cascade Mountains (Anthony et al. 2006). Similarly, California spotted owls inhabit relatively mesic conifer-dominated forests at higher elevations in the Sierra Nevada, but also inhabit oak (Quercus spp.) woodlands in the foothills and conifer-hardwood mixtures in the Peninsular Mountains of southern California (Verner et al. 1992). And Mexican spotted owls inhabit both rocky canyonlands and a range of forest types from wet mixed-conifer forests to drier pine-oak forest types (USDI FWS 2012). There is little reason to expect fire behavior to be similar across the range of occupied forest types, and therefore may be little reason to expect owl response to be similar among subspecies, geographic areas, and forest types. In this context, note that most of the available studies are on California spotted owls (Table 1). Far less information is available from the range of the northern and, especially, the Mexican spotted owl, and the responses of these subspecies may not be well described by studies of California spotted owls.

Some of the variation among studies also may stem from local variability among the fires themselves. Of particular interest here are recent studies by Lee and Bond (2015 a) and Jones et al. (2016) in the same general area within the Sierra Nevada. Both studies evaluated occupancy rates of California spotted owls one year after large “megafires” (defined as fires that burned ≥10 000 ha; Stephens et al. 2014 b) that burned in 2013 (Rim Fire; Lee and Bond 2015 a) or 2014 (King Fire; Jones et al. 2016) and that included areas that burned with high severity. In the Rim Fire, Lee and Bond (2015 a) concluded that owls were not negatively impacted by high-severity wildfire because occupancy rates for spotted owls were higher than previously published rates reported in either burned or long-unburned areas, and the amount of high-severity fire in designated owl Protected Activity Centers (Verner et al. 1992)did not affect pair occupancy. In contrast, Jones et al. (2016) concluded that high-severity wildfire negatively impacted owls and that megafires posed a threat to spotted owls because occupancy probability for spotted owls declined by 22 % the year after the King Fire and declined by almost nine-fold in sites that burned at >50 % high severity.

Thus, studies of California spotted owls in two megafires in the same general area and time frame generated opposite conclusions regarding owl response to high-severity wildfire. Jones et al. (2016) hypothesized that the difference in results and conclusions reflected differences in the spatial pattern of areas that burned at high severity within each fire. High-severity patches were larger and more contiguous in the King Fire than in the Rim Fire (Figure 1) and overlapped numerous owl territories (Jones et al. 2016: figure 2). This explanation suggests that owl response is driven by the structure of the landscape mosaic that remains following wildfire. Where fires create a landscape with adequate amounts of remnant older forests, owls likely will be able to occupy that landscape (Bond 2016). In contrast, fires that create large patches that burn with high severity and leave inadequate amounts of remnant older forests are likely to render those areas unsuited for occupancy by spotted owls (Jones et al. 2016).

Figure 1
figure 1

Maps showing fire severity within two recent megafires in the Sierra Nevada, within the range of the California spotted owl. The King Fire (left) burned 40 106 ha in 2014, with 17 419 ha burned at high severity (43.4 %). The Rim Fire (right) burned 104 080 ha in 2013, with 20 698 ha burned at high severity (19.9 %). Fire severity data obtained from USGS MTBS Project (2016). Scale is equal for both maps. The box in the lower left corner indicates the location of the fire within California.

Full size image
Figure 2
figure 2

A severely burned area 14 years after the Rodeo-Chediski megafire, Arizona. Note that many snags remain standing within drainages, but in more exposed areas most snags have already fallen. As the remaining snags fall, large openings devoid of the elevated perches required by foraging owls will be created.

Full size image

Most Studies Do Not Address Long-Term Response to High-Severity Wildfire

Most studies of spotted owls in burned areas occurred less than four years post fire and therefore evaluated only short-term responses (Table 1). Although some studies of California spotted owls included longer maximum post-fire periods, most of those studies included multiple fires that burned in different years (Roberts et al. 2011; Lee et al. 2012, 2013; Lee and Bond 2015 b), with occupancy surveys conducted over the same years for all fires. Thus, for any given survey year, time since fire varied among fires. Results in these studies were pooled across fires, time since fire was not explicitly evaluated, and the relative proportion of data analyzed by post-fire period was not specified. Consequently, most data available on occupancy and use of burned areas by spotted owls is limited to the first few years post fire, and data showing that severely burned areas continue to be occupied by spotted owls over longer time frames range from sparse to nonexistent across subspecies (Table 1). Because spotted owls have relatively long lifespans and exhibit high site fidelity (Bond et al. 2002, Blakesley et al. 2006, Ganey et al. 2014 a), continued short-term occupancy of burned areas could be partly due to site fidelity of pre-fire residents.

Studies of Occupancy Rates May Not Be Sufficient to Evaluate Owl Response

In addition to covering only short time frames, many studies of owl response to wildfire evaluated only occupancy rate or sometimes occupancy and reproduction (Table 1). Although these studies are valuable, those parameters alone are not sufficient to evaluate owl response. Only one study has reported on long-term demographic rates of marked owls following fire (Rockweit et al. 2017). Results again varied among fires evaluated. However, burned territories in some fires showed higher turnover in resident owls than unburned territories, but remained occupied over time despite that high turnover. The authors hypothesized that these burned territories functioned as population sinks with continued occupancy supported by immigration from nearby unburned source territories. This suggests that longer-term empirical data, using parameters that measure demographic rates of marked owls beyond territory occupancy and reproduction, are required to understand the effects of high-severity fire on spotted owls over a time frame relevant to population persistence.

The Duration of Post-Fire Positive Effects on Foraging Habitat is Unknown

Spotted owls are known to forage in burned areas, including severely burned areas. The principal benefit to spotted owls from high-severity fire is hypothesized to be improved foraging habitat due to increases in small mammal abundance following fire (Bond et al. 2009, Bond 2016). There is evidence that some small mammals are more abundant following fire (Converse et al. 2006, Fontaine and Kennedy 2012, Ganey et al. 2014 a), but results are not entirely consistent across taxa (Fontaine and Kennedy 2012, Roberts et al. 2015) and most studies of post-fire small mammal communities do not track abundance over long post-fire periods. A review by Fontaine and Kennedy (2012) found no studies of small mammals that tracked response for periods >4 yr post fire. Thus, for species of small mammals that show post-fire increases in abundance, the persistence of that increase is unknown.

The length of time during which spotted owls can forage effectively in severely burned areas also is unknown. Spotted owls hunt from elevated perches, and salvage logging in severely burned areas can quickly render large areas deficient of the elevated perches required for effective foraging (Bond 2016). Burned trees also fall quickly in some areas, however, so even unlogged areas may soon become deficient in elevated perches. For example, many unlogged areas in the Rodeo-Chediski megafire, Arizona, were largely devoid of standing snags by 14 years post fire (Figure 2). Elsewhere in Arizona ponderosa pine (Pinus ponderosa Laws) forests, 41 % of snags fell within seven years following a wildfire (Chambers and Mast 2005), and coarse woody debris peaked between 6 yr and 12 yr post fire in two studies across chronosequences following wildfire (Passovoy and Fulé 2006, Roccaforte et al. 2012). This suggests that, at least in some areas or forest types, foraging habitat created by high-severity fire may be useful only for a short time. Ultimately, as snags fall, areas that are burned but not logged will provide the same low-quality foraging habitat reported by Bond (2016) for areas that are logged after fire. Note that salvage logging in spotted owl foraging habitat will hasten this process, however, and therefore still has negative consequences for spotted owl foraging habitat (Bond 2016).

Potential Improvements in Foraging Habitat Do Not Compensate for Loss of Nesting Habitat

Spotted owls selectively nest in forests featuring large trees and high canopy cover throughout their range (Forsman et al. 1984; Grubb et al. 1997; Hershey et al. 1998; May et al. 2004; Blakesley et al. 2005; Ganey et al. 2013, 2016; Tempel et al. 2014, 2016). Where such forests are available for nesting, owls will forage in a wide variety of forest and even non-forest cover types or edges (Call et al. 1992; Ward et al. 1998; Ganey et al. 1999, 2003; Comfort et al. 2016). Consequently, nesting habitat for spotted owls generally is considered more limited in amount and distribution than foraging habitat (USDI FWS 2012). As noted above, increase or improvement in owl foraging habitat from high-severity fire may be ephemeral, whereas the loss of nesting habitat is long-term, because regrowth of stands of large old trees with high canopy cover can take >100 yr. Thus, short-term gains in foraging habitat due to high-severity wildfire are unlikely to offset the loss of nesting habitat in the longer term.

High-Severity Wildfire is Sufficiently Widespread to Constitute a Potential Threat

Proponents of the opinion that fire is not a threat to spotted owls may underestimate the potential extent and impact of high-severity wildfire. Bond (2016) argued that the number of owl territories subject to detrimental amounts of high-severity fire were likely to be small in any wildfire. This conclusion, however, may not be supported for megafires, which are large by definition and often include extensive areas that burn with high severity. The extent and severity of megafires has increased in recent decades (Miller et al. 2009, USDI FWS 2012) and they are predicted to likely increase as climate warms (McKenzie et al. 2004, Westerling et al. 2006, Littell et al. 2009), and impacts to owls and their habitat may be considerable (Stephens et al. 2016).

To assess the extent of megafires within the range of the spotted owl, we searched data from Monitoring Trends in Burn Severity (USGS MTBS Project 2016), a multi-year project designed to map fire perimeters and burn severity across all lands since 1984 (Eidenshink et al. 2007). We restricted our search to wildfires >10 000 ha (after Stephens et al. 2014 b) occurring on US Forest Service lands between 2000 and 2014 in the states of Arizona, California, New Mexico, Oregon, and Washington. The time period thus includes only recent fires, and roughly coincides with the period included in most studies of spotted owls and fire. By restricting the search to US Forest Service lands, we focused on areas most likely to harbor spotted owls. We did not include fires in Utah and Colorado, within the range of the Mexican spotted owl, because many owls in these areas occur in rocky can-yonland habitat that may be less directly affected by fire (Willey and van Riper 2007, USDI FWS 2012). We also used fire vicinity maps from USGS MTBS Project (2016) to eliminate fires outside of the range of the spotted owl within the states searched (i.e., east of the Cascade Mountains or Sierra Nevada), or in lowland valley areas unlikely to contain spotted owl territories.

Although our search was conservative and underestimated fire extent within the range of the spotted owl, we identified 105 megafires matching our search criteria (Supplemental Table 1, available from the senior author, lists the megafires used to develop the summary data in Tables 2 and 3, along with associated attribute data and data documentation.). Fires >100 000 ha occurred in the last 15 years within the ranges of all three owl subspecies, and individual fires burned >50 000 ha at high severity within the range of all three subspecies (Table 2). These fires appear to have the potential to impact considerable numbers of owl territories, even within a single fire. This could place significant stress on local owl populations, particularly where owls occur in small insular populations within the range of California and Mexican spotted owls.

Table 2 Mean and maximum area (ha) of “megafires” within the ranges of three subspecies of spotted owls between 1 January 2000 and 31 December 2014, and mean and maximum area burned at high severitya.
Full size table
Table 3 Cumulative area burned and area burned at high severity (ha) in “megafires” within the ranges of three subspecies of spotted owls, between 1 January 2000 and 31 December 2014a.
Full size table

The cumulative effect of fire on spotted owls and their habitat is more detrimental than the effects of any single fire, however. The megafires identified in our search collectively burned 3 567 518 ha, including 723 319 ha that burned with high severity (Table 3). We do not know how much of that area met conditions for spotted owl habitat; our point here is simply that these fires cumulatively burned an immense area and thus have the potential to impact considerable numbers of owl territories in a relatively short time period. These fires also created relatively large patches that burned with high severity (e.g., Figure 1).

The above analysis estimated overall area burned and was not restricted to the types of forest used for nesting by spotted owls. In a more focused analysis, Stephens et al. (2016) estimated area burned in potential California spotted owl nesting habitat in the Sierra Nevada and concluded that high-severity wildfire posed a significant threat to persistence of California spotted owl nesting habitat. Their analysis indicated that, between 2000 and 2014, 85 046 ha of potential spotted owl nesting habitat was burned by wildfires, which resulted in ≥50 % basal area mortality and reduced canopy cover to >25 %. Area burned increased over the period from 1970 to 2014 (Stephens et al. 2016: figure 4), and a regression model based on that trend predicted that the cumulative amount of potential nesting habitat burned at >50 % tree basal area mortality would exceed the amount of existing nesting habitat within 75 years (Stephens et al. 2016: figure 5). Similarly, USDI FWS (2012) concluded that megafires were the biggest threat facing Mexican spotted owls, based on the potential for rapid cumulative loss of nesting habitat to high-severity wildfire.

Forest Restoration Treatments Warrant Further Study

Forest restoration treatments have been proposed as a means to reduce fuels and fire risk, increase forest resiliency, and restore natural fire regimes (Covington and Moore 1994, Allen et al. 2002, Brown et al. 2004, Stanturf et al. 2014, Hagmann et al. 2017). There is empirical evidence that such treatments can modify fire severity, at least in drier forest types (Raymond and Peterson 2005, Prichard et al. 2010, Waltz et al. 2014, Roccaforte et al. 2015, Ziegler et al. 2017), and some (but not all) studies modeling fire behavior concluded that such treatments could modify both fire severity and extent, thus reducing the risk of loss of nesting habitat to high-severity wildfire (Lee and Irwin 2005, Ager et al. 2007, Cushman et al. 2011, Chiono et al. 2017; but see Odion et al. 2014, Baker 2015 b, Hanson and Odion 2016). These treatments likely are unnecessary in areas where owls inhabit wetter forests characterized by long fire-free periods (Agee 1993)

Existing studies on the effects of fuels reduction treatments on spotted owls universally suggest negative effects from these treatments (Meiman et al. 2003, Seamans and Gutiérrez 2007, Stephens et al. 2014 a, Tempel et al. 2014). These studies, however, are few in number, the mechanisms underlying owl response remain unclear, only short-term responses have been studied, and a limited range of treatment types have been evaluated. There may be important tradeoffs between short-term impacts due to treatments and long-term benefits from those treatments due to reduction in the risk of high-severity fire (Lee and Irwin 2005, Ager et al. 2007, Tempel et al. 2015, Chiono et al. 2017). There also may be types of treatments that have not yet been evaluated that could reduce fire risk while maintaining habitat conditions suitable for spotted owls. Consequently, further studies aimed at understanding the effects of various types of restoration treatments on spotted owls are badly needed, as well as simulations of habitat trajectories with and without forest treatments (Lee and Irwin 2005, Ager et al. 2007, Odion et al. 2014, Tempel et al. 2015, Chiono et al. 2017).

Conclusions

Considerable uncertainty remains about the responses of spotted owls to wildfire, especially responses to high-severity fire over longer time frames and across subspecies. Relatively few studies have evaluated the effects of wildfires on spotted owls, especially for the Mexican and northern subspecies, and results of existing studies are short-term and sometimes contradictory. To be clear, we do not assume that all wildfires are detrimental to spotted owls, and in fact surmise that low-severity and moderate-severity wildfires pose little threat to spotted owls. Available evidence suggests that high-severity wildfire can be detrimental, however, depending on spatial pattern and extent. The considerable recent extent of high-severity wildfire within the ranges of these subspecies (Table 3; see also Stephens et al. 2016), coupled with the trend toward increasing extent and severity of megafires (Miller et al. 2009, Stephens et al. 2016), suggests that the cumulative effects of these fires could be significant throughout the range of all three subspecies of the spotted owl.

Targeted research studies are needed to evaluate response of spotted owls to high-severity wildfire under different conditions of landscape pattern and in different forest types. Those studies should occur across a variety of forest types within the ranges of all three subspecies, should incorporate fires differing in size and extent and spatial pattern of high-severity fire, and should occur in the absence of salvage logging to avoid confounding fire effects with logging effects. We recommend studies of marked owls to provide information on the demography of populations within burned areas, rather than relying on estimates of occupancy rates (Tyre et al. 2001).

Forest restoration treatments may aid in reducing fire risk and habitat loss in some situations, particularly in drier forest types (Waltz et al. 2014, Roccaforte et al. 2015, Chiono et al. 2017, Ziegler et al. 2017). We acknowledge that not all such treatments will be beneficial to owls and their habitat. Implementing restoration treatments in and around owl habitat, therefore, is not without risk (USDI FWS 2012), and we urge caution in that implementation. Treatments should be located strategically based on models of fire behavior and spread to optimize gains in reduction of fire risk relative to area treated (Ager et al. 2007, 2010; Finney et al. 2007). Locating treatments outside of owl nesting habitat could reduce landscape-scale fire risk in proximity to owl nesting habitat while minimizing short-term impacts to such habitat (USDI FWS 2012; Stephens et al. 2014 a, 2016; Tempel et al. 2015; Jones et al. 2016). When fire models indicate that treatments are needed within occupied owl habitat to reduce fire risk, treatments differing in intensity and spatial extent should be tested, and their effects on spotted owls should be carefully monitored, with the goal of identifying treatment types that successfully reduce fire risk while retaining habitat conditions suitable for spotted owls. Evaluating owl response to a range of treatment types that vary in extent and intensity, simulating extent and spatial pattern of habitat remaining in landscapes with and without such treatments, and coupling these models with demographic simulation models (e.g., Landguth and Cushman 2010), will help us understand tradeoffs between wildfire and restoration treatments with respect to conserving spotted owls and their habitat.

We also suggest that managers consider wider use of managed fire (both wild and prescribed) in the context of forest restoration (Collins et al. 2011, Hunter et al. 2011, Parks et al. 2014, North et al. 2015, Stephens et al. 2016, Huffman et al. 2017). Again, this is not without risk. Introducing or allowing fire to burn in owl habitat that features high fuel loads presents special challenges for fire managers. Nonetheless, allowing fire to burn in these areas when weather conditions are favorable (i.e., cool and wet with calm winds) could reduce the risk of high-severity fire while maintaining some of the habitat elements essential to spotted owls, as opposed to waiting for such areas to burn under the extremely hot, dry, and windy conditions characteristic of most megafires. Relative to mechanical treatments, managed fire is likely to be far more economical and may result in landscapes of greater spatial complexity preferred by spotted owls (Collins et al. 2011, Larson and Churchill 2012, Comfort et al. 2016). As with both wildfires and mechanical treatments, the effects of managed fire on owls and their habitat should be carefully evaluated.

The debate over spotted owls, wildfire, and forest restoration has important implications for management strategies over large forested landscapes, as well as for how we integrate conservation of threatened and endangered species and their habitats with other resource management objectives. This debate also involves numerous complicated issues. We do not pretend to resolve this debate here, but we hope that this paper stimulates productive dialogue among wildlife biologists, forest ecologists, and fire ecologists, and spurs additional research on historical fire regimes and wildfire and treatment effects on spotted owls and their habitat. Until better information is available on such effects, we argue that it is premature to conclude that high-severity wildfire poses no threat to spotted owls, or to dismiss restoration treatments as a tool in reducing fire risk and habitat loss.