Conservation Evidence strives to be as useful to conservationists as possible. Please take our survey to help the team improve our resource.

Providing evidence to improve practice

Action: Use prescribed fire or modifications to burning regime in forests Amphibian Conservation

Key messages

  • Eight of 14 studies (including three randomized, replicated, controlled studies) in Australia, North America and the USA found no effect of prescribed forest fires on amphibian abundance or numbers of species. Four found that forest fires had mixed effects on amphibian abundance depending on species, species and year or season of burn. Three found that fires increased amphibian abundance or numbers of species. One found that abundance decreased with fires.
  • Two studies (including one randomized, replicated, controlled study) in the USA found that numbers of amphibian species and abundance increased or abundance decreased with time since prescribed forest fires.
  • One before-and-after study in the USA found that spotted salamander hatching success increased following a prescribed forest fire.


Supporting evidence from individual studies


A controlled, site comparison study in 1982–1983 of sandhill-scrub habitat in west central Florida, USA (Mushinsky 1985) found that controlled burns resulted in higher species diversity and abundance of amphibians. The 7-year burn cycle plot had the greatest number of species in both years (7-year cycle: 16–20; 2-year: 10–15; 1-year: 14–16; unburned: 10–15). Although burn plots had a greater fluctuations in species diversity over the two years than the unburned plot, numbers of captures were higher. Captures tended to be highest in 7- and 1-year burn plots (7 years: 115–307; 2 years: 102–187; 1 year: 126–203; unburned: 71–125). The 1-year cycle was most consistent for supporting high numbers of individuals and species. A 1 ha plot was established for each burn cycle in adjacent strips. These were compared to a plot unburned for 20 years. Burns were in May–June. Five drift-fence arrays with pitfall traps and an artificial cover board were established/plot. Traps were checked 5–6 times/week in April–October 1983–1984.



A replicated, controlled, site comparison study in 1994 of native forest and managed near Brisbane, Australia (Hannah & Smith 1995) found that prescribed fires in native forest resulted in increased amphibian abundance but not species richness. In native forest there was a significantly higher number captured in 5-year burn cycles than unburned sites (5-year cycle: 127; 3-year: 85; unburned: 51). In plantations, numbers were similar (burned seven years ago: 37; burned two years ago: 48; unburned: 39). There was no significant difference in species richness between treatments (native: 3–4; plantation: 6). Treatments in native forest (1.5 ha; two replicates) were: burned in autumn–winter on a 3-year cycle (burned 1991), in winter–spring on a 5-year cycle (burned 1993) or unburned (since 1973). In the plantation (25 ha) treatments were: burned two or seven years ago or unburned. Drift-fencing with pitfall traps and active searching were used for monitoring in January or March 1994 (75–180 trap nights/treatment).




A controlled study in 1992–1993 of pine stands in Maryland, USA (McLeod & Gates 1998) found that annual prescribed burns resulted in significantly lower amphibian abundance. Captures were significantly lower in the burned compared to unburned stand for total amphibians (74 vs 391), salamanders (8 vs 105), ranid species (6 vs 20) and frogs and toads (66 vs 214). The same was true for two of 10 frog and toad species, adults of two of four salamander species and young of the year for three frog species. The other species showed no significant difference between treatments. Study sites were an unburned mixed pine-hardwood stand (5 ha) and a pine stand (4 ha) that had been burned annually since 1981, with alternating thirds being burned from 1988. Monitoring was undertaken using three drift-fences with pitfall and funnel traps per site in March–July 1992–1993.



A replicated, controlled study in 1995–1996 in a national forest in Carolina, USA (Ford et al. 1999) found that prescribed fires did not tend to affect the abundance of salamanders. There were no significant difference in numbers of blue ridge two-line salamanders Eurycea wilderae, Jordan's salamanders Plethodon jordani or mountain dusky salamanders Desmognathus ochrophaeus captured in burned and unburned areas. Seepage salamander Desmognathus aeneus captures were significantly lower in the riparian zone of the burned compared to unburned areas in 1996 (0.2 vs 1.3). Monitoring was undertaken for two weeks immediately before an April burn and after the burn in June 1995 and August 1996 at two sites. Drift-fencing with pitfalls and snap-traps were installed at three locations in the upper slope, mid-slope and riparian zone at each site. Visual searches were also undertaken. An unburned area at one of the sites was monitored in the same way.



A randomized, replicated, controlled study in 1997–1998 of pine sandhills in Florida, USA (Litt et al. 2001) found that prescribed burning resulted in similar or lower abundance of amphibians compared to unburned sites. In 1997 there was no significant difference between treatments for any species. In 1998, capture rates were significantly lower in prescribed burn plots and herbicide understory removal plots than fire suppressed (control) plots for southern toad Bufo terrestris (burn: 0; understory: 0.002; no burn: 0.008; reference: 0.003 captures/trap days). Capture rates did not differ between burned, understory removal or fire suppressed treatments for oak toad Bufo quercicus or eastern narrowmouthed toad Gastrophryne carolinensis. In 1997 (not 1998), similarity indices indicated that burned plots were significantly more similar to reference (frequently burned) sites than understory removal or fire suppressed plots (burn: 0.76; understory: 0.49; no burn: 0.49). Treatments were in randomly assigned 81 ha plots within four replicate blocks in spring 1997. Data were also collected from four frequently burned reference sites. Monitoring was with drift-fencing and pitfall traps in April–August 1997–1998.



A replicated, site comparison study in 1994–1996 of mature pine forest in Georgia, USA (Miller et al. 2001) found that there was no apparent difference between amphibian abundance or numbers of species in forest burned in the growing or dormant season. Total amphibian captures and numbers of species were similar between plots burned in the growing season (abundance: 32; species: 7) and dormant season (abundance: 19; species: 4). Captures were higher in unburned hardwood forest (abundance: 101; species: 14). Sample sizes were considered too small for statistical analysis. Three plots burned in the 1994 growing season (April–August; 3-year cycle) and three burned in the dormant season (January–March) were selected. Three adjacent hardwood plots were also surveyed. Three drift-fences with 12 pitfall traps and four artificial cover boards were installed within each plot. Monitoring was undertaken over four weeks, four times in 1995–1996.



A replicated, controlled study in 2001 of a bottomland hardwood forest in Georgia, USA (Moseley, Castleberry & Schweitzer 2003) found that prescribed burning did not increase the  abundance, diversity or richness of amphibians. Abundance did not differ significantly at burned and unburned sites for all amphibians (burned: 43; unburned: 62), salamanders (2 vs 6) or frogs and toads (39 vs 50). The same was true for species richness overall (burned: 8; unburned: 8 species), for salamanders (2 vs 2) or frogs and toads (6 vs 6). The volume of coarse woody debris was similar in burned (60 m3/ha) and unburned stands (128 m3/ha). Amphibians were monitored in three winter-burned and unburned stands from July to October 2001. Drift-fencing with pitfall traps, artificial cover boards and PVC pipe refugia were randomly placed within each site.



A review in 2003 of the effects of prescribed fire on amphibians in North America (Pilliod et al. 2003) found that results were mixed. Four studies found that amphibian abundance or abundances of some species were lower in burned compared to unburned stands. One study found that abundance of certain species was higher following burning, two found mixed results depending on species and two found no significant differences between treatments. One of two studies found that species richness was greatest in 5–7 year burn cycles and the other found no difference between burned and unburned stands. The majority of studies focused on short-term responses (1–3 years post-burn), with only one of ten investigating longer-term effects (five years post-burn).



A site comparison study of 15 ponds in a pine forest in South Carolina, USA (Schurbon & Fauth 2003) found that amphibian abundance and species richness increased with time since prescribed burns. Abundance of all amphibians and frogs and toads increased significantly with time since burning. This was not the case for salamanders. Amphibian species richness also increased significantly over time following burns. This was likely to be because salamanders were rarely encountered at sites burned within two years, but became more abundant with time. Amphibians were monitored at 15 ponds with five different prescribed burn (in winter/spring) histories: 0, 1, 3, 5 and 12 years after burns. Drift-fences, tree-frog shelters, calling censuses, minnow trapping and visual surveys were used.



A randomized, replicated, controlled study in 1995–1996 of shelterwood-harvested oak stands in Virginia, USA (Keyser et al. 2004) found that prescribed burns did not affect amphibian abundance or species richness. There were no significant differences in relative abundances between burned and unburned sites for all amphibians (burned: 10–15; unburned: 6), eastern red-backed salamanders Plethodon cinereus (7–11 vs 3) or American toads Bufo americanus (3 vs 2). Amphibian species richness did not differ significantly between burned and unburned sites (2–3 vs 5). Three replicates (2–5 ha) of four randomly assigned treatments were applied in 1995: burning in February, April or August, or unburned. Three uncut reference sites were also monitored. Amphibians were monitored using pitfall traps (20/site) for 53 nights in June, July and October 1996.



A replicated, controlled study in 2003–2004 of pine savanna in Mississippi, USA (Langford et al. 2007) found that prescribed burning resulted in a greater abundance but similar diversity of amphibians compared to unburned sites. Greater numbers of amphibians were found at burned than unburned sites (275 vs 90). However, species diversity was similar (burned: 13; unburned: 10). Some species were significantly more abundant in burned compared to unburned areas including oak toads Bufo quercicus (125 vs 9) and southern leopard frogs Rana utricularia (51 vs 2). In comparison, a small number of species were more common in unburned sites including the pig frog Rana grylio (13 vs 2). A low intensity burn was undertaken over a large proportion of a National Wildlife Refuge in 2003. From January to June 2004, amphibians were monitored at three burned and three unburned sites. Visual encounter surveys (200 m transects), minnow traps (six/site) and PVC tubes (five/site) were used.



A randomized, replicated, controlled study in 2001–2004 in hardwood forest in Carolina, USA (Greenberg & Waldrop 2008) found that prescribed burns did not increase overall amphibian abundance or species richness, but did increase abundance of frogs and toads. The relative abundance of total amphibians, salamanders and green frog Rana clamitans did not differ significantly between treatments. However, abundances of anurans (frogs and toads) and American toads Bufo americanus were significantly higher in burn treatments compared to controls (anurans: 52–54 vs 8; American toads: 50 vs 10 captured/100 nights). Species richness did not differ significantly (burned: 5; burned with understory reduction: 5; control: 3). There were three 14 ha replicates of each randomly assigned treatment: prescribed burn, burn and mechanical understory reduction and controls. Understory reduction was undertaken in winter 2001–2002 and burns in March 2003. Drift-fences with pitfall and funnel traps were used for monitoring in August–October 2001 and May–September 2002–2004.



A replicated, site comparison study in 1999–2001 of pine woodland in western Arkansas, USA (Perry, Rudolph & Thill 2009) found that controlled burning did not affect amphibian species abundance. There was no significant difference between numbers of captures in burned and unburned plots for all amphibians (73 vs 59), all frogs and toads (71 vs 55), individual species or salamanders (2 vs 4). The most abundantly caught species, the western slimy salamander Plethodon albagula, was captured almost exclusively in unmanaged woodland (28 of 29 captures). Nine plots (11–42 ha) that had been thinned (1980–1990) and then burned at least three times at 3–5-year intervals were sampled. These were compared to three unmanaged, unburned plots. Controlled fires were in March–April. Three drift-fence arrays with pitfall and box traps were established/plot. Traps were checked weekly in April-September 1999–2001.



A before-and-after study in 2005–2007 of a pond in restored mixed forest in Illinois, USA (Sacerdote & King 2009) found that prescribed burning resulted in increased hatching success for spotted salamander Ambystoma maculatum. Eggs failed to hatch in 2005, but following burning, hatching success of egg masses was 29% in 2006 and 53% in 2007. Restoration started in 2000 and included destruction of drainage tiles, clearing of invasive plants, prescribed burning and removal of leaf litter. The burn was in autumn 2005. An egg mass was placed in two mesh enclosures (56 x 36 x 36 cm) in the pond. Eggs were monitored every five days until hatching was complete.



A controlled, before-and-after study in 2001–2006 of ponderosa pine forest in Idaho, USA (Arkle & Pilliod 2010) found that a prescribed fire had no significant effect on the density of rocky tailed frog tadpoles Ascaphus montanus. During the study, the density of tadpoles decreased by 50% in both burned (pre-burn: 2.3; post-burn: 1.1/m2) and unburned catchments (pre: 2.7; post: 1.6). A prescribed burn was undertaken in May 2004 and burned 12% of one catchment. Four nearby unburned catchments were monitored for comparison. Tadpoles were monitored using kick-sampling in 30 transects (1 m wide) per stream in 2001–2006.



A replicated, before-and-after study in 1988–2008 of 25 wetlands in forest and grassland reserves in Indiana, USA (Brodman 2010) found that the relative abundance of salamanders declined following prescribed spring, but not autumn or winter burns. The six forest species declined significantly (82–100%) following spring burns and took an average of five years to recover to pre-burn levels. Declines were not associated with autumn or winter burns and tiger salamander Ambystoma tigrinum and eastern newt Notophthalmus viridescens increased at two sites after an autumn burn. Monitoring was undertaken the year before and after burns. Each site was visited monthly for three months in spring and one in summer or autumn. Visual searches, minnow traps, dipnets and seines were used to survey entire small ponds (< 0.25 ha) and 50 m of adjacent upland habitat, or along transects for larger ponds.



A replicated, controlled before-and-after study in 2001–2007 of hardwood forest in West Virginia, USA (Ford et al. 2010) found that although population responses were difficult to interpret following two prescribed fires, results suggested that there was no significant affect on the salamander assemblage. Mountain dusky salamanders Desmognathus ochrophaeus and red-backed salamander Plethodon cinereus counts were greater following burns compared to before burns or unburned controls. However, authors considered that this was due to increased use of artificial cover boards in response to reduced leaf litter following fires. Treatments were burn plots on upper slopes or lower slopes (n = 20), half of which were fenced and control plots that were unburned and unfenced (n = 4). Burns were in 2002–2003 and 2005. Cover board arrays were used to monitor salamanders before and after two fires in April-October in 2001–2007.



A randomized, replicated study in 1999–2001 of nine restored pine woodlands in western Arkansas, USA (Perry, Rudolph & Thill 2012) found that overall numbers of amphibians were highest in the first year after burns compared to the following two years. This was true for total amphibians (1st year: 114; 2nd year: 53; 3rd year: 51/stand) and anurans (1st: 112; 2nd: 51; 3rd: 49). However, this trend was largely due to high numbers of dwarf American toads Bufo americanus charlessmithi in the first year (83 vs 27–31). Fowler’s toads Bufo fowleri were also captured most often in year one stands (2.0 vs 0.1–0.2). Salamanders captures did not differ between years after burn. In 1999–2001, stands (11–42 ha) were burned on a 3-year cycle, so three were burned each year in March–April. Stands had been thinned at least nine years previously and had undergone 3–7 prescribed burns at 2–5 year intervals. Monitoring was undertaken using three drift-fence arrays per stand (15 m) connected to central funnel traps in April–September in 1999–2001.



Referenced papers

Please cite as:

Smith, R.K., Meredith, H. & Sutherland, W.J. (2018) Amphibian Conservation. Pages 9-65 in: W.J. Sutherland, L.V. Dicks, N. Ockendon, S.O. Petrovan & R.K. Smith (eds) What Works in Conservation 2018. Open Book Publishers, Cambridge, UK.