Meta‐analysis of avian and small‐mammal response to fire severity and fire surrogate treatments in U.S. fire‐prone forests

May combine two claims.

Thinning plus low/moderate-severity fire relative to low-severity fire alone provided a greater level of positive response, with 52% of 27 species having larger effect sizes.

“Thinning plus low/moderate-severity fire relative to lowseverity fire alone provided a greater level of positive response (52% of 27 species had larger effect sizes, based on 95% confidence interval overlap; Table 3).”

May combine two claims.

The literature was most abundant for low/moderate- and high-severity fire at 0-4 years, with more than 10 studies in each category, followed by thinning studies.

“The literature was most abundant for low/moderate-and high-severity fire at 0-4 years, with .10studies in each category; this was followed by studies of thinning (Table 1).”

May combine two claims.

Cassin's Finch and Chipping Sparrow, previously classified as mixed responders, showed positive responses to low/moderate-severity fire and neutral responses to high-severity fire in this meta-analysis.

“Similarly, Cassin's Finch and Chipping Sparrow were previously classified as mixed responders, whereas we found positive responses to low/moderateseverity and neutral responses to high-severity fire for both species.”

May combine two claims.

For all species pooled, low/moderate-severity fire produced largely neutral responses (53%), with negative responses rare (19%) and positive responses intermediate (28%).

“Treatment responses (all species pooled; Fig. 2) suggest largely neutral responses to low/ moderate-severity fire or thinning alone (53%, 50% respectively; Fig. 3b), with negative responses being the most rare (19%, 13%, respectively; Fig. 3f ), and levels of positive response intermediate (28%, 38% respectively; Figs. 2 and 3).”

May combine two claims.

High-severity fire data were restricted almost entirely to birds (60 species), with data for only one small mammal (deer mouse).

“Data for high-severity fire were restricted almost entirely to birds (N ¼ 60 species; data existed for just for one small-mammal species (deer mouse; Tables 1 and 2 and Appendix B).”

May combine two claims.

For all species pooled, thinning alone produced largely neutral responses (50%), with negative responses rare (13%) and positive responses intermediate (38%).

“Treatment responses (all species pooled; Fig. 2) suggest largely neutral responses to low/ moderate-severity fire or thinning alone (53%, 50% respectively; Fig. 3b), with negative responses being the most rare (19%, 13%, respectively; Fig. 3f ), and levels of positive response intermediate (28%, 38% respectively; Figs. 2 and 3).”

May combine two claims.

No studies of large mammals met inclusion criteria; only one study on medium-sized mammals and one on bats were found.

“Data for large mammals (e.g., mesocarnivores, ungulates) that met our inclusion criteria did not exist, only one study on medium-sized mammals and one study on bats were found (Table 1; Appendix B)”

May combine two claims.

The meta-analysis located 41 studies covering 119 bird species and 17 small-mammal taxa that assessed short-term (up to 4 years) responses to thinning, low/moderate-severity fire, high-severity fire, and thinning plus prescribed fire; other taxa and longer periods had too little data for quantitative synthesis.

“We found 41 studies of 119 bird and 17 smallmammal species that examined short-term responses ( 4 years) to thinning, low/moderateand high-severity fire, and thinning plus prescribed fire; data on other taxa and at longer time scales were too sparse to permit quantitative assessment.”

May combine two claims.

Across all 27 species, there was little evidence of a negative interaction between thinning and low/moderate-severity fire; responses appeared additive or changed from neutral to negative/positive.

“Additionally, across all 27 species, we found little evidence of a negative interaction of thinning and low/ moderate-severity fire where responses changed from positive to negative or vice versa; instead, responses appeared to be additive in nature (e.g., Western Bluebird; Fig. 3e) or a neutral to negative/positive change (e.g., golden-mantled ground squirrel, Ovenbird; Fig. 3a, c).”

May combine two claims.

Data for reptiles and amphibians were identified but restricted to low/moderate-severity fire and fire surrogate treatments for few species, so they were excluded; high-severity data at longer time scales were insufficient and also excluded.

“Some data for reptiles and amphibians were identified (Table 1), but these were restricted to low/moderate-severity fire and fire surrogate treatments for a small number of species and thus were not included; data for high severity at longer time scales were insufficient and also were not included (Table 1; Appendix B).”

May combine two claims.

The combination of thinning plus low/moderate-severity fire was characterized by higher positive response (41%) and decreased neutral response (34%).

“The combination of thinning plus low/ moderate-severity fire was characterized by higher levels of positive response (41%) and decreased neutral response (34%; Figs. 2 and 3a).”

May combine two claims.

Low/moderate-severity fire had the most data (126 species: 109 birds and 17 small-mammal taxa), while thinning had 48 species and thinning plus low/moderate-severity fire had 29 species.

“Data for low/moderate-severity fire were the most abundant (N ¼ 126 species; Fig. 2) and included 109 bird species and 17 small-mammal taxa (Table 2), whereas data for thinning or thinning plus low/moderate-severity fire were more limited (N ¼ 48 species for thinning; N ¼ 29 species for thinning plus low/moderate-severity fire; Table 2, Fig. 2).”

May combine two claims.

For species with data in more than one treatment, low/moderate-severity fire and thinning showed similar response patterns for 47% of species, while 40% showed a larger positive response to thinning relative to low/moderate-severity fire.

“For species where data were available for more than one treatment, comparison of low/moderate-severity fire and thinning suggested a similar response pattern (47% of species had no difference in response; Fig. 3d, e) and 40% showed a larger positive response to thinning relative to low/moderate-severity fire (Table 3).”

May combine two claims.

In the authors' qualitative examination of guild responses, no strong trends were found because each guild included diverse species.

“Similar limitations apply to the use of nesting and foraging guilds for bird species; we qualitatively examined guild responses in our own data (J.B. Fontaine and P. L. Kennedy, unpublished data) and found no strong trends, owing to the diverse set of species included within each group.”

May combine two claims.

Brown Creeper, previously classified as a negative responder to fire by Kotliar et al. (2002), showed a positive response to low/moderate-severity fire and a neutral (negative point estimate) response to high-severity fire in this meta-analysis.

“For example, Brown Creeper was classified by Kotliar et al. (2002) as a negative responder to fire; in contrast, we found a positive response to low/moderate-severity fire and a neutral (negative point estimate) response to highseverity fire.”

• cites Kotliar et al. 2002Not checked yet

  Not checked in the free run.

May combine two claims.

The golden-mantled ground squirrel showed neutral responses to low/moderate-severity fire or thinning alone but a positive response to the combination.

“The golden-mantled ground squirrel (Fig. 3a) typified this pattern of response, with neutral responses to low/ moderate-severity fire or thinning alone, but a positive response to the combination of the two treatments.”

May combine two claims.

Many bird species responded negatively to high-severity fire, predominantly canopy-nesting and foliage-foraging species, but many others did not.

“Although many bird species responded negatively to high-severity fire (predominantly canopy-nesting and foliage-foraging species), many others did not.”

May combine two claims.

Most low/moderate-severity fires in the data were prescribed and occurred at smaller spatial extents, while most high-severity fires were wildland fires burning at larger spatial extents.

“This is particularly problematic for comparisons of low/moderate-severity fire and high-severity fire, where most low/ moderate severity fires were prescribed and occurred at smaller spatial extents, and most high-severity fires were wildland fires burning at larger spatial extents.”

May combine two claims.

Some species, such as deer mouse and Townsend's Solitaire, do not respond to low/moderate-severity fire but respond positively to high-severity fire, or respond more positively to high-severity than to low/moderate-severity fire.

“Some species such as deer mouse or Townsend's Solitaire do not respond to low/moderate-severity fire, but positively respond to high-severity fire (or respond more positively to high-severity fire than to low/moderate-severity fire).”

May combine two claims.

Comparisons in Table 3, based on species common across all three treatments, show little difference from overall pairwise comparisons, suggesting the bias from pooling species was negligible.

“However, the comparisons in Table 3, based on species in common across all three fire and fire surrogate treatments (low/moderate-severity fire, thinning, thinning plus low/moderate-severity fire) show little difference from the overall pairwise comparisons, suggesting that this source of bias was negligible.”

May combine two claims.

Too few studies adjusted counts or vital rates for detection or capture probabilities, so all studies were treated identically regarding potential detection/capture bias.

“Ultimately, too few studies adjusted counts or vital rates to account for detection or capture probabilities; thus all studies were treated identically in terms of their potential detection or capture bias”

• AssumptionRests on an assumption: Detection or capture bias is similar enough across studies and treatments that treating all studies identically does not bias treatment comparisons.

May combine two claims.

The study supports earlier single-study-area conclusions that there is no single wildlife response to wildland fire or fire management; instead, responses vary by species.

“With regard to individual species, our study provides strong support for conclusions of earlier studies conducted in single study areas that there is no such thing as ''a wildlife response'' to wildland fire or fire management activity.Instead, responses of wildlife to fire and fire surrogate treatments vary by species (Kotliar et al. 2007).”

• cites Kotliar et al. 2007Not checked yet

  Not checked in the free run.

May combine two claims.

Because most landscapes support specialists requiring early-seral or older forests and species requiring multiple seral stages, a mosaic of successional stages is likely needed to maximize landscape biodiversity.

“Given that most landscapes support specialists that require either earlyseral or older forests, as well as species that require multiple seral stages throughout their life cycle, it is likely that a mosaic of successional stages is needed in the landscape to maximize its biodiversity (Fontaine et al. 2009, Roberts et al. 2010).”

• cites Fontaine et al. 2009Not checked yet

  Not checked in the free run.

• cites Roberts et al. 2010Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

A random effect of study was included to account for typical noise and true unmeasured differences among studies.

“We included a random effect of study in the analysis to account for both typical ''noise'' as well as true unmeasured differences among studies, a frequent issue in ecological studies (Rosenberg et al. 2000).”

• cites Rosenberg et al. 2000Not checked yet

  Not checked in the free run.

May combine two claims.

Vegetation regeneration, fuel dynamics, and nutrient cycling differ substantially between fire and thinning.

“regeneration of vegetation, fuel dynamics, and nutrient cycling following fire vs. thinning do differ substantially (Boerner et al. 2009)”

• cites Boerner et al. 2009Not checked yet

  Not checked in the free run.

May combine two claims.

Few generalities exist about faunal responses to fire even though knowledge is increasing rapidly.

“Thus, although we are in a period of rapidly increasing knowledge, few generalities exist with regard to faunal responses to fire.”

May combine two claims.

Vanderwel et al. (2007) examined bird response to thinning across a range of tree retention levels and found much greater negative response than this study did.

“Vanderwel et al. (2007) examined bird response to thinning across a range of tree retention levels and found much greater levels of negative response than we did.”

• cites Vanderwel et al. 2007Not checked yet

  Not checked in the free run.

May combine two claims.

Most previous reviews of fire effects on fauna used qualitative vote-counting and made little formal assessment of effect sizes for fire severities, fire surrogates, and time since fire.

“To date, most reviews have relied on qualitative comparisons, employing a vote-counting approach (Rosenberg et al. 2000) to the literature and making little formal assessment of the effect size of fire severities, fire surrogates, and time since fire on fauna.”

• cites Rosenberg et al. 2000Not checked yet

  Not checked in the free run.

May combine two claims.

Because observations were pooled across forest types and wildland/prescribed fire studies, inference is limited to the stand scale (about 10-50 ha); landscape-scale effects could not be examined.

“Therefore the appropriate inference is to the stand scale (;10-50 ha); variation due to landscapescale effects of wildland fire, such as quantity of burned habitat, could not be examined.”

May combine two claims.

Fire-created early-successional habitat in the western United States is very limited and rare compared with historic levels and may limit several management-relevant species.

“supports the assertion made by Noss et al. (2006) that fire-generated early-successional habitat in the western United States is extremely limited and rare relative to historic levels and may be a limiting factor for a number of management-relevant species.”

• cites Noss et al. 2006Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

Using thinning instead of prescribed fire may be justified for wildlife, especially where prescribed fire is difficult to implement.

“Thus, our results suggest that use of thinning in lieu of prescribed fire may be warranted for wildlife, particularly in areas where implementation of prescribed fire is problematic (e.g., peri-urban areas).”

May combine two claims.

In northwestern California, Northern Spotted Owl has higher survival and reproductive rates within heterogeneous stands of mature mixed-evergreen forest originating from fire, despite lower occupancy and lower survival immediately following high-severity fire in the same region.

“In northwestern California, the threatened Northern Spotted Owl (S. o. caurina) has higher survival and reproductive rates within heterogeneous stands of mature mixed-evergreen forest originating from fire (Franklin et al. 2000), despite lower occupancy and lower survival immediately following high-severity fire in the same region (Clark 2007).”

• cites Franklin et al. 2000Not checked yet

  Not checked in the free run.

• cites Clark 2007Not checked yet

  Not checked in the free run.

May combine two claims.

Some variation in the study's results may be attributed to pooling across forest types and historic fire regimes, but the impact of this potential confounding variation cannot be evaluated.

“However, some of the variation in our results may be attributed to these two factors, but we cannot evaluate the impact of this potential source of confounding variation on our results.”

May combine two claims.

Comparison of low/moderate- and high-severity fires was restricted to western North America because stand-replacement fires generally do not occur in southeastern U.S. forests, except the Florida scrub pine ecosystem for which no published studies met criteria.

“Comparison of low/moderate-and high-severity fires was restricted to western North America because standreplacement fires generally do not occur in southeastern U.S. forests (an important exception to this is the Florida scrub pine, Pinus clausa, ecosystem; it has a stand-replacement fire regime, but we found no published studies meeting our criteria from this ecosystem).”

May combine two claims.

Even for high-severity fire, data were limited and related to several publications of longitudinal studies of the same fire.

“However, even for high-severity fire, data were limited and related to several publications of longitudinal studies of the same fire (Raphael et al. 1987).”

• cites Raphael et al. 1987Not checked yet

  Not checked in the free run.

May combine two claims.

The study's scope of inference is limited to the stand scale and mostly to birds and some ground-dwelling small mammals.

“Our scope of inference is limited to the stand scale (mostly due to the operational nature of prescribed fire and thinning studies), and mostly to birds and some ground-dwelling small mammals.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The effect of repeated fires and fire interval on wildlife cannot be addressed by this study.

“Another important management question that cannot be addressed by our study is the effect of repeated fires and fire interval on wildlife.”

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

Wildland fire use works well in some places, but in many other areas it is problematic.

“As Fule et al. (2004) observe, wildland fire use works well in some places, but in many other areas it is problematic.”

• cites Fule et al. 2004Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Combined with temporal variation from forest succession after disturbance, mixed-severity fire has the potential to create landscape heterogeneity that could maximize availability of wildlife habitat to the broadest range of species.

“Combined with the temporal variation provided by forest succession after disturbance, mixedseverity fire has the potential to create the landscape heterogeneity that could maximize availability of wildlife habitat to the broadest range of species (Fontaine et al. 2009).”

• cites Fontaine et al. 2009Not checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: The spatial arrangement of extant forest and stand-replacement patches across stand-to-watershed scales creates a mosaic that supports more species than any single cover type.

May combine two claims.

Data to examine the influence of time since fire or treatment were sparse and concentrated in studies of high-severity fire.

“Data to examine the influence of time since fire or treatment were sparse and concentrated in studies of high-severity fire.”

May combine two claims.

High-severity fire may be an important process in some forest ecosystems, probably providing habitat conditions absent from unburned forests or forests subjected only to low-severity fire, but present in forests that historically experienced mixed-severity fire.

“This suggests that high-severity fire may be an important process in some forest ecosystems; it probably provides habitat conditions not found in unburned forests or forest subjected to low-severity fire, but those that experienced some level of mixed-severity fire historically (e.g., in many western forest types, and Great Lakes states).”

Probably not a substantial claim: may not pass S3 (load-bearing).

Because the data were at stand scale, fire severity was treated coarsely as either low/moderate or high severity.

“Due to the standscale nature of our data, fire severity was treated somewhat coarsely as either low/moderate or high severity.”

May combine two claims.

In a meta-analysis, positive and negative responses where 95% confidence intervals do not overlap zero reflect consistency among studies and allow inference to ecological response.

“In a meta-analysis such as this study, positive and negative responses (where 95% confidence intervals of the effect size do not overlap zero) reflect consistency among studies and allow for inference to the ecological response of an organism.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Previous studies documented similar results for gradients of thinning intensity.

“Previous studies have also documented similar results for gradients of thinning intensity (Hayes et al. 2003).”

• cites Hayes et al. 2003Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The study could not formally test the response types (additivity, threshold, dominance).

“We could not test these response types formally and suggest that investigators treat these as alternative hypotheses to be tested in future studies.”

May combine two claims.

The patterns found were consistent with additivity of treatments, threshold responses, or one treatment overwhelming the other.

“Consistent with Hurteau et al. (2008), the patterns we found were consistent with additivity of treatments, threshold responses, or response to one treatment overwhelming that of the other.”

• cites Hurteau et al. 2008Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

The study's results support the growing literature indicating that there is no one-size-fits-all prescription for incorporating disturbance into land management.

“Our results support the growing literature (Fuhlendorf et al. 2006, Fontaine et al. 2009) indicating that there is no one-size-fits-all prescription when it comes to incorporating disturbance into land management.”

• cites Fuhlendorf et al. 2006Not checked yet

  Not checked in the free run.

• cites Fontaine et al. 2009Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The importance of surface fires in maintaining fire-dependent species has been well documented.

“The importance of surface fires in maintaining fire-dependent species has been well documented (Engstrom et al. 1984), but the effect of variance in fire interval is uncertain; repeated high-severity fires, particularly at shorter intervals that may occur in regions with rapid postfire vegetation recovery (Thompson et al. 2007), are not well understood.”

• cites Engstrom et al. 1984Not checked yet

  Not checked in the free run.

• cites Thompson et al. 2007Not checked yet

  Not checked in the free run.

May combine two claims.

In the short term, high-severity fire may negatively impact Northern Spotted Owls, but may create suitable habitat at about 20 years and high-quality habitat by 60-80 years postfire.

“Thus, in the short term, high-severity fire may negatively impact Northern Spotted Owls, but may create suitable habitat at ;20 years and high-quality habitat by 60-80 years postfire.”

Classified as probably not a substantial claim.

Underlying bias may confound treatment comparisons if the species pool varies in composition and thus life history and other attributes.

“Underlying bias may confound these comparisons if the species pool varies in its composition and thus life history and other attributes.”

Classified as probably not a substantial claim.

Comparisons across treatments required pooling across species.

“Similarly, comparisons across treatments (i.e., Fig. 2, Table 3) required us to pool across species.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The importance and widespread occurrence of mixed-severity wildfire is an emerging perspective.

“An emerging perspective (Donato et al. 2009b, Thompson and Spies 2009, Halofsky et al. 2011) is the importance and widespread occurrence of mixed-severity wildfire.”

• cites Donato et al. 2009bNot checked yet

  Not checked in the free run.

• cites Thompson and Spies 2009Not checked yet

  Not checked in the free run.

• cites Halofsky et al. 2011Not checked yet

  Not checked in the free run.