Background information and definitions
Prescribed burning is a practice used in forest management where controlled burns are conducted to reduce the risk of more damaging uncontrolled natural fires and to stimulate tree germination. Controlled burning alters forest structure, opens up the tree canopy and creates potential roosts in snags.
Although prescribed or controlled burning may benefit bats, there may also be negative effects such as heat injury, smoke and carbon monoxide poisoning, and arousal from torpor. Consideration must be given to fire intensity, ignition procedures and seasonal timing of burns (Dickinson et al. 2010).
Dickinson M.B., Norris J.C., Bova A.S., Kremens R.L., Young V. & Lacki M.J. (2010) Effects of wildland fire smoke on a tree-roosting bat: integrating a plume model, field measurements, and mammalian dose-response relationships. Canadian Journal of Forest Research, 40, 2187–2203.
Supporting evidence from individual studies
A replicated study in 2003–2004 in a deciduous forest in Missouri, USA (Boyles & Aubrey 2006) found that evening bats Nycticeius humeralis roosted only in areas of the forest where prescribed burning had occurred. Twenty-three bats were tracked to 63 tree roosts in burned areas, and no roosts were found in unburned areas. The burned area of the forest had a more open canopy and more dead trees than the unburned area. Prescribed burning began in 1999 after 50 years of fire suppression and was done every two years in March or April in 55% of the study area. Bats were caught from March 2003 to March 2004 using mist nets across forest roads between the burned and unburned areas of the 1,200-ha forest and in 2–3 ponds or roads in both areas. Twenty-three bats (11 females and 12 males) were fitted with radio-transmitters and tracked to roost trees each day until the transmitter was shed or expired.Study and other actions tested
A replicated, controlled, site comparison study in 2001–2002 in nine pine forest sites in South Carolina, USA (Loeb & Waldrop 2008) found that burned tree stands had similar activity of three bat species to unburned control tree stands, and two bat species had higher activity in thinned stands than burned stands. Activity of big brown bats Eptesicus fuscus, eastern red bats Lasiurus borealis and eastern pipistrelles Perimyotis subflavus did not differ significantly between burned tree stands (big brown bats: average 0.3 bat passes/night; red bats: 0.3 bat passes/night; eastern pipistrelles: 0.1 bat passes/night) and unburned control stands (big brown bats: 0.1 bat passes/night; red bats: 0.5 bat passes/night; eastern pipistrelles: 0.1 bat passes/night). Activity was higher in thinned tree stands than burned stands for big brown bats (1.2 bat passes/night) and eastern red bats (0.7 bat passes/night), but similar for eastern pipistrelles (0.4 bat passes/night). Nine 14 ha stands (loblolly pine Pinus taeda and shortleaf pine Pinus echinata) were surveyed with three replicates of three treatment types: thinning to an average of 576 live trees/ha (in winter 2000–2001), prescribed burning (burned in April 2001 with strip head fire and flanking fires, average 532 live trees/ha), and a control with no treatment (average 755 live trees/ha). Bat activity was sampled with two bat detectors at random points in each of 12 stands for two full nights/month from May–August 2001 and 2002.Study and other actions tested
A replicated, controlled, site comparison study in 2007–2008 of four mixed forest sites in West Virginia, USA (Johnson et al 2009) found female northern myotis bats Myotis septentrionalis roosting in tree stands treated with prescribed fire and in unburned forest, and roost switching frequency and the distance between roost trees did not differ between burned and unburned forest. Twenty-five roosts were in burned tree stands and 44 in unburned forest, but the difference was not tested for statistical significance. Roost switching frequency and the distance between roost trees did not differ significantly between burned (1–6 days, average 152 m) and unburned forest (1–5 days, 230 m). In April–May 2007 and 2008, three stands (45, 13 and 21 ha) were burned for one day using a strip head fire technique. The remainder of the 1,900-ha forest was left unburned. Bats were captured over streams, pools, trails, and service roads at burned and unburned sites using mist nets in May–August 2007 and 2008. In 2007, three female bats were radio-tracked to eight roosts. In 2008, 33 female bats were radio-tracked to 65 roosts, four of which were used previously in 2007.Study and other actions tested
A replicated, controlled, before-and-after study in 2006–2007 of three mixed forest sites in Kentucky, USA (Lacki et al 2009) found that burned forest had more female northern myotis Myotis septentrionalis roosts than unburned forest, and home ranges were closer to burned than unburned forest. Following prescribed fires, more female northern myotis bat roosts were in burned forest (26 roosts, 74%) than unburned forest (nine roosts, 26%), although no statistical tests were carried out. The average size of home ranges and core areas did not vary significantly between bats radio-tracked before (home range: 60 ha; core area: 11 ha) and after fires (home range: 72 ha; core area 14 ha), but home ranges were closer to burned habitats than unburned habitats following fires. Two sites (435 ha, 185 ha) that were previously unburned were subject to prescribed burning in April 2007, with 54% of the area burned. A third site (2,400 ha) was left unburned. Bats were captured in June–July 2006 and April–September 2007 using mist nets over ponds in burned and unburned sites. Eighteen female bats were radio-tracked nightly for an average of six days.Study and other actions tested
A replicated, controlled, site comparison study in 2007–2008 in two mixed forests in West Virginia, USA (Johnson et al 2010) found male Indiana bats Myotis sodalis roosting in tree stands treated with prescribed fire and in unburned forest, and roost switching frequency and the distance between roost trees did not differ between burned and unburned forest. Sixteen roosts were in burned tree stands and 34 roosts were in unburned forest, but the difference was not tested for statistical significance. Roost switching frequency and the distance between roost trees did not differ significantly between burned (1–4 days, average 220 m) and unburned forest (1–2 days, 477 m). In April–May 2007–2009 three stands (12, 13 and 121 ha) within one of two of the forests (Fernow Experimental Forest, 1900 ha) were subjected to prescribed burning using a strip head fire technique. In the other forest (Petit Farm, 400 ha) in March 2003 an escaped campfire had burned part of the forest stand. Control stands were unburned areas in each of the two forests. Bats were captured over streams, pools, ponds and trails using mist nets in burned and unburned forest in June–July 2004–2006 at Petit Farm and June–July 2008–2009 at Fernow Experimental Forest, and also at a cave swarming site at Fernow Experimental Forest in September–October 2007–2008. A total of fifteen male bats were radio-tracked.Study and other actions tested
A site comparison study in 2004–2005 in nine forest fragments within the Chicago metropolitan area, USA (Smith & Gehrt 2010) found that two of seven forest fragments that had undergone restortation, including prescribed burning, had higher bat activity than two unrestored forest fragments. Bat activity was higher in two forest fragments that had been restored with multiple prescribed burns, invasive plant species removal and snag recruitment (average 7–19 bat passes/survey) than in two control sites with no restoration (average 1–4 bat passes/survey). Bat activity was similar between control sites and five other forest fragments that had been restored with multiple prescribed burns and various combinations of invasive species removal, snag recruitment and deer population control (1–6 bat passes/survey). Six bat species were recorded in total (see original paper for data for individual species). The study does not distinguish between the effects of prescribed burning and the other interventions carried out. Fire suppression over the last 100 years had altered the structure of the nine forest fragments (10–260 ha in size). Seven of the nine forest fragments were being restored to open up the canopy, reduce tree density and remove invasive plant species. At each of nine sites, four bat detectors recorded bat activity for 4 h from sunset for five nights/year in June–September 2004 and May–August 2005.Study and other actions tested
A replicated, site comparison study in 2008–2009 of 48 sites in two mixed forests in Florida, USA (Armitage & Ober 2012) found that the frequency of prescribed burns had no effect on the activity of cluttered habitat bat species, but open habitat bat species were recorded less in forest sites with longer periods between burns. The activity of bat species adapted to cluttered habitats did not differ significantly in forest sites with different burn frequencies (1–2 year burn: average 6 bat passes/site/night; 3–5 year burn: 4 bat passes/site/night; >8 year burn: 6 bat passes/site/night). Bat species adapted to open habitats had lower activity in forest sites with a longer period between burns (>8 years: 0.03 bat passes/site/night) than forest sites with more frequent burns (1–2 year burn: 0.1 bat passes/site/night; 3–5 year burn: 0.05 bat passes/site/night). Twenty-four 40 ha study plots were randomly selected in each of two forests with eight plots for each of three burn frequencies (burns every 1–2 years, 3–5 years or >8 years). Bat detectors recorded nightly bat activity at two randomly chosen sites/burn frequency/night for four evenings/week with sites rotated weekly in May–August 2008 and 2009.Study and other actions tested
A replicated, controlled, before-and-after, paired sites study in 2008 in six tropical eucalypt woodland sites in northern Queensland, Australia (Inkster-Draper et al 2013) found that prescribed burning resulted in higher bat activity and a change in species composition. A greater number of bat calls were recorded at treatment sites after prescribed burning (average 2,423) than before (1,174). There was no significant difference in bat calls at unburned control sites over the same period (‘before’: 1,008; ‘after’: 1,568 bat calls). Species composition also differed at the treatment sites before and after burning but did not differ at unburned control sites over the same period (data reported as statistical model results). At least 10 bat species were recorded (see original paper for data for individual species). One site from each of six pairs was burned with a low intensity fire for two days (treatment) with the other remaining unburned (control). Bat activity was recorded using bat detectors at six paired sites for 336 bat detector nights in June and July 2008 before burning, and for 234 bat detector nights during August, September and October 2008 after burning.Study and other actions tested
A replicated, controlled, site comparison study in 2010–2012 in 26 savanna and woodland sites in Missouri, USA (Starbuck et al 2015) found that prescribed burning increased occupancy rates of burned sites for one of five bat species. Occupancy rates of the evening bat Nycticeius humeralis increased at burned sites with a greater number of prescribed fires in the past 10 years (data reported as statistical model results). The number of fires did not have a significant effect on the occupancy rates of burned sites for four other bat species (northern long-eared bat Myotis septentrionalis, big brown bat Eptesicus fuscus, eastern red bat Lasiurus borealis, tri-coloured bat Perimyotis subflavus). At each of 26 sites, surveys were carried out at 4–28 sampling points in managed forest and unmanaged mature forest. Managed forests had been burned (with 1–8 fires over 10 years) and thinned to restore savanna or woodland. Bat detectors recorded bat activity at each sampling point for two full consecutive nights during 1–2 years in 2010–2012.Study and other actions tested
A replicated, randomized, site comparison study in 2013–2014 of 10 hardwood tree stands in Tennessee, USA (Cox et al 2016) found that burned and thinned tree stands had higher overall bat activity for three of four treatment types than untreated tree stands. Overall bat activity was higher in tree stands burned in the autumn and thinned to 7m2/ha (average 280 bat passes) or 14m2/ha (292 bat passes) than untreated control tree stands (95 bat passes). Tree stands burned in the spring had higher bat activity than controls stands when thinned to 14m2/ha (656 bat passes) but not 7m2/ha (123 bat passes). Six groups of bat species were recorded (see original paper for data for individual species groups). The study does not distinguish between the effects of burning and thinning. Each of four treatments (burning in the autumn or spring with thinning to 7 or 14 m2/ha) was randomly applied to two tree stands (20 ha, 80–100 years old). Two tree stands were untreated controls (average 20 m2/ha). Overstorey thinning was carried out in June 2008 and prescribed fires in October 2010 and 2012 (autumn) and March 2011 and 2013 (spring). Each of 10 stands was surveyed with a bat detector for seven full nights on three occasions in May–July 2013 and 2014.Study and other actions tested
A replicated, site comparison study in 2006–2010 of twelve tree stands in two upland hardwood forests in Ohio, USA (Silvis et al 2016) found that burned and thinned tree stands had higher overall bat activity than untreated tree stands. Overall bat activity was higher in tree stands burned and thinned with 50% of the overstorey retained (average 16–30 bat passes/night) and tree stands burned and thinned with 70% of the overstorey retained (14–24 bat passes/night) than in untreated control stands (3–4 bat passes/night). Four bat species or species groups were recorded (see original paper for data for individual species). The study does not distinguish between the effects of burning and thinning. In each of two forests, four tree stands (10 ha) were treated with thinning (commercially thinned between June 2005 and March 2006 with 50% or 70% overstorey retained) and prescribed fire (backing and strip fires in autumn 2009 or spring 2010) and two tree stands were untreated controls. In each of 12 tree stands, eight points were sampled with bat detectors for 3 h/night over a total of six nights in May–August 2006 and June–September 2009 and 2010.Study and other actions tested
A replicated, before-and-after, controlled study in 2015–2016 of two upland pine Pinus elliottii forests in south Florida, USA (Braun de Torrez et al 2018) found that prescribed burning increased the activity of Florida bonneted bats Eumops floridanus. Activity of Florida bonneted bats was higher at treatment sites after prescribed burning than before but did not change at unburned control sites (data reported as statistical model results). Similarly, the activity of Florida bonneted bats did not differ between treatment and control sites before burning but was higher at burned than unburned control sites after prescribed burning. Two prescribed burns (one in the dry season, one in the wet season) were carried out in each of two upland pine forest conservation areas. Burned areas were 46–549 ha. Bat activity was recorded at three sites within each of four treatment areas and at three adjacent unburned control sites for 12 nights before burning and 24 nights after burning in February and March 2016 (wet season) and June 2015 and July 2016 (dry season).Study and other actions tested
A replicated, paired sites study in 2014–2015 in 11 paired areas of mixed forest in Tennessee and Kentucky, USA (Burns et al 2019) found that burned sites had a higher occupancy of five bat species or species groups than unburned sites, and burn severity had a negative effect on Myotis spp. and tri-colored bats Perimyotis subflavus. Overall, burned sites had a higher occupancy of five bat species/species groups (big brown bats Eptesicus fuscus/silver-haired bats Lasionycteris noctivagans, eastern red bats Lasiurus borealis/evening bats Nycticeius humeralis, Myotis spp., tri-colored bats, hoary bats Lasiurus cinereus) than unburned sites (data reported as statistical model results). Occupancy of Myotis spp. and tri-colored bats was higher in sites burned with a low severity than a moderate severity. A total of 164 paired forest sites were surveyed within 11 burned and 11 unburned areas (each 10–1,150 ha). Burned areas had been treated with low or moderate intensity prescribed fire (or wildfire at one site) at least once in the past 10 years. Bat presence and activity were recorded simultaneously with bat detectors at paired sites for at least two full nights in May–August 2014 and 2015.Study and other actions tested
A replicated, site comparison study in 2009–2011 in managed upland forest in Kentucky, USA (Johnson et al 2019) found that home ranges of female Rafinesque’s big-eared bats Corynorhinus rafinesquii were similar distances from burned and unburned forest, whereas the home ranges of male bats were closer to unburned forest. Home ranges of female bats (including non-reproductive, pregnant, lactating, and post-lactating females) were not located significantly closer to burned forest than unburned forest (data reported as statistical model results). Home ranges of male bats were further from burned forest than unburned forest. Prescribed burns of low-moderate intensity were carried out each spring within an area of deciduous, coniferous, and mixed forest. Burned areas were an average of 252 ha. Forty-one adult Rafinesque’s big-eared bats (33 females, eight males) from four colonies were captured in mist nets and radio-tagged. Each bat was tracked to an average of 38 locations over 2–5 nights in May–September 2009–2011.Study and other actions tested
A replicated, randomized, site comparison study in 1990–2004 of 18 forest sites in New South Wales, Australia (Law et al 2019) found that unlogged forest blocks burned every four years had higher overall bat activity and more bat species than blocks burned every two years or unburned blocks, but species composition did not differ and burning had no effect in logged forest. Higher bat activity (2–3 times) and more bat species were recorded in unlogged forest blocks burned every four years than in blocks burned every two years or unburned blocks (data reported as statistical model results). Species composition did not differ significantly between treatments. Burning every two or four years in logged forest blocks did not have a significant effect on bat activity, species richness or composition compared to unburned logged forest. Fifteen bat species or species groups were recorded (see original paper for data for individual species). Three blocks (8–56 ha) in logged and unlogged forest were randomly allocated to one of three treatments: burning every four years, burning every two years, no burning. Prescribed burns began in 1990. Logging was carried out in 1987–1988 with mature trees retained (33% of the original basal area). Each of 18 blocks was surveyed with two bat detectors for two consecutive nights in December 1999, 2001 and 2004.Study and other actions tested