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: Retain riparian buffer strips during timber harvest Amphibian Conservation

Key messages

  • Twelve studies investigated the effectiveness of retaining buffer strips during timber harvest for amphibians.
  • Six replicated and/or controlled studies in Canada and the USA compared amphibian numbers following clearcutting with or without riparian buffer strips. Five found mixed effects on abundance depending on species and buffer width. One found that amphibian abundance was significantly higher with buffers.
  • Eleven studies, including 10 replicated and/or controlled studies in Canada and the USA and one meta-analysis, compared amphibian numbers in forest with riparian buffers retained during harvest to unharvested forest. Six found mixed effects depending on species or volume of existing downed wood. Four found that abundance and species composition were similar to unharvested forest. Two found that numbers of species and abundance were lower than in unharvested forest.
  • Two of four replicated studies (including one randomized, controlled, before-and-after study) in Canada and the USA found that numbers of amphibian species and abundance were greater in wider riparian buffer strips. Two found that there was no difference in abundance in buffers of different widths.


Supporting evidence from individual studies


A controlled, before-and-after study in 1988–1991 at three hardwood forest sites in Oregon, USA (Cole et al. 1997) found that the effects of retaining riparian buffer zones on amphibians were unclear. Three of six species showed no changes in capture rates after total clearcutting and no significant differences in captures in riparian buffers and upslope areas (rough-skin newts Taricha granulosa, Dunn's salamanders Plethodon dunni and red-legged frogs Rana aurora). Capture rates of ensatinas Ensatinae schscholtzii and Pacific giant salamanders Dicamptodon tenebrosus decreased after clearcutting and tended to be lower in buffers than upslope. Western redback salamanders Plethodon vehiculum increased the first year after logging and then decreased. Herbicide treatment had no effect on species. Each site had plots (>8 ha) with each treatment: unharvested control; clearcut and broadcast burned; and clearcut, broadcast burned and sprayed with herbicide (1.3 kg/ha). Clearcuts had 20 m wide untreated riparian buffer strips. Cut sites were planted with fir seedlings. Amphibians were monitored one year before and for two years after treatments using pitfall trapping. Traps were checked daily for eight days in dry and wet seasons.



A replicated, site comparison study in 1994–1995 along streams at 29 forest sites in western Oregon, USA (Vesely 1997, Vesely & McComb 2002) found that clearcut forest with retained riparian buffers had significantly higher amphibian density than total clearcut plots (12 vs 6/1,000 m2). However, compared to unharvested sites clearcut sites with riparian buffers had significantly lower total salamander abundance (21 vs 30) and species richness (3 vs 5) and abundance of three individual salamander species. Two species did not differ between treatments. Overall and individual species density did not differ significantly within plots with riparian buffers and unharvested sites (amphibians: 12 vs 13/1,000 m2). Amphibian density was significantly higher within wide (>40 m) compared to narrow (<20 m) buffers (13 vs 5/1,000 m2). The same was true for species richness (5 vs 2). Seventeen clearcut sites (< 5 years old) with riparian buffers (0–64 m wide) and 12 unharvested sites (> 100 years) were selected. Visual encounter surveys were undertaken in three 20 x 40 m streamside plots/site (within buffers, clearcut, unharvested areas) in April–May and November–December 1994 and March–May 1995.



A replicated, controlled, before-and-after study in 1996–1998 of a mixed wood forest in Alberta, Canada (Hannon et al. 2002) found that forest buffers of 20–200 m around lakes maintained amphibian abundance for three years after harvest. Abundance was not significantly different before and after harvest, within or between buffer widths, or compared to unharvested areas and protected forests. Species composition did not change after harvest. Four lakes were selected in three regions and were assigned to buffer strip treatments of 20, 100 or 200 m wide, or were controls within protected forest. Clearcuts were 2–49 ha, with two to four cuts around each lake in 1996, the remainder was left unharvested. Amphibians were monitored using groups of three pitfall traps at 40 m intervals within sampling grids (400 x 100 m) parallel to lakes. Sampling was undertaken in May–June and July–August 1996–1998, for 5–8 days/lake each season.



A randomized, replicated, controlled, before-and-after study in 2000–2003 of forest in Maine, USA (Perkins, Malcolm & Hunter 2006) found that amphibian abundance tended to be higher when riparian buffers were retained during harvest. Captures were significantly higher with 11 m and 23 m buffers for American toads Bufo americanus (clearcut: 0.6; 11 m buffer: 1.0; 23 m buffer: 3.4; unharvested: 0.5/100 trap nights) and wood frogs Rana sylvatica (clearcut: 0.8; 11 m: 1.4; 23 m: 2.0; unharvested: 2.2). Red-backed salamanders Plethodon cinereus did not differ (0.1–0.3). In forest cut 4–10 years previously, captures of wood frog and American toads were also significantly higher in buffers than clearcuts. Red-backed salamanders showed a similar trend. However, abundance of salamanders and frogs were significantly or tended to be lower in buffers than unharvested forest. Fifteen headwater streams were randomly assigned to 6 ha treatments: clearcut with buffers of 0, 11 or 23 m wide, partial harvest (23–53%) or unharvested. Monitoring was undertaken using drift fences with pitfall traps and visual surveys in June–September, one year before and two years after harvesting. Twelve sites harvested 4–10 years earlier were also monitored in one year. Treatments were: clearcutting with 23–35 m buffers, partial harvest and unharvested (> 50 years).



A replicated, controlled, before-and-after study in 1995–2002 of amphibians in managed forest stands at 11 sites in Oregon, USA (Olson & Rugger 2007) found that retaining riparian buffers maintained amphibian abundance in the first two years after tree thinning. There was no significant decrease in four species within buffers following thinning (change: −0.1 to 0.1 animals/m2). Rough-skinned newts Taricha granulosa and coastal giant salamander Dicamptodon tenebrosus numbers increased within buffers following thinning (0.007–0.034/m2) and declined at unthinned control sites (−0.043 to 0.008/m2). Forty-five streams were assigned riparian buffers of 6, 15, 70 or 145 m on each side within tree thinning areas (from 600 to 200 trees/ha). Thinning took place in 1997–2000. Monitoring was undertaken in spring and summer, before treatment, in 1995–1999 and for two years after treatment, in 1998–2002. Amphibians were sampled in 10 units/stream using hand sampling, electrofishing and visual counts of bank sides (2 m wide). Twenty-three streams within unharvested areas were also monitored.



A controlled, before-and-after, site comparison study in 1998–2001 at two forest sites in western Oregon, USA (Rundio & Olson 2007) found that the amount of pre-existing downed wood affected the response of salamanders to forest thinning with riparian buffers. At the site with high volumes of existing downed wood, there was no significant change in amphibian capture rates following thinning with three different buffer widths. However, at the site with little downed wood, capture rates declined following thinning with buffers of ≥6 m or ≥15 m, but not ≥70 m. At the two sites, treatments were unharvested or thinned (to 200 trees/ha; 10% cut in groups; 10% patches retained; deadwood retained) with riparian buffer widths of ≥6 m (streamside-retention), ≥15 m (variable-width) or ≥70 m. Monitoring was undertaken in May–June before and two years after thinning. Visual count surveys were along 102 m transects perpendicular to each stream bank (7–8/treatment).



A replicated, site comparison study in 2005 of three forest sites in Oregon, USA (Kluber, Olson & Puettmann 2008) found that there was no significant difference between amphibian captures in riparian buffers and unharvested forest 5–6 years after harvest. Captures did not differ significantly between thinned and unharvested, or between two buffer widths (6 and >15 m) for all amphibians, western red-backed salamanders Plethodon vehiculum or ensatina Ensatina eschscholtzii. However, captures did decrease significantly with distance from stream for all amphibians and red-backed salamanders. Captures varied with distance for ensatina. Overall, 60% of captures occurred within 15m of the stream. Each 12–24 ha site had two streams within forest that had been thinned (600 to 200 trees/ha) with riparian buffers (6 m or >15 m wide) retained in 2000 and one stream with no harvesting. Amphibians were sampled by visual counts once in April-June within five 5 x 10 m plots at four distances from each stream (up to 35 m).



A replicated, controlled study in 2005–2007 of salamanders in five headwater streams in North Carolina, USA (Peterman & Semlitsch 2009) found that retaining 30 m riparian buffers during timber harvest maintained salamander populations. Two-lined salamander Eurycea wilderae larvae were significantly more abundant within 30 m buffers (413 larvae) and unharvested streams (171–533) than in streams with 9 m or no buffers (72–73). However, black-bellied salamanders Desmognathus quadramaculatus showed no difference in abundance between treatments (25–34 larvae). Treatments were timber harvest with riparian buffers of 0, 9 or 30 m retained on both sides of the stream. The two controls were no harvest. Timber was harvested in 2005–2006. Salamanders were monitored within three 40 m sampling blocks along streams in May–August 2006 (9 m buffer and controls) and 2007 (all sites). Animals were captured using 48 leaf litter bags/site each 1–2 weeks.



A randomized, replicated, controlled study in 2003–2005 of 11 forest ponds in east-central Maine, USA (Veysey, Babbitt & Cooper 2009) found that the impact of buffer zones on spotted salamander Ambystoma maculatum migration behaviour depended on weather conditions. Migration rate and distance of salamanders from ponds did not differ significantly between treatments. However, the probability of migration differed significantly between the 100 m buffer and unharvested, but not 30 m buffer treatments. If rainfall was low, salamanders were more likely to move in the 100 m compared to unharvested treatment, above 390 mm of cumulative rainfall the opposite was true. Ponds were randomly assigned to treatments: clearcut with 30 m or 100 m buffers or unharvested. Concentric 100 m wide clearcuts were created around buffers surrounding ponds in 2003–2004. Salamanders were captured in pitfall traps along drift-fences as they left breeding ponds in spring. Forty salamanders were radio-tracked (6–21/treatment) in April–November 2004–2005.



A meta-analysis of global studies of amphibians in harvested forests (Marczak et al. 2010) found that riparian buffers were not effective at maintaining amphibian abundance. Amphibian abundance was significantly lower in buffers compared to unharvested areas. Frogs and toads (15 studies) showed greater differences between buffers and unharvested sites (both positive and negative) compared to salamanders (16 studies). There was no significant effect of buffer width or time since buffer establishment on the size of the difference in abundance between buffers and unharvested sites (amphibians, birds, small mammals and arthropods combined). Wider buffers did not result in greater similarity between buffer and unharvested sites. A meta-analysis was undertaken using published data from 31 studies comparing abundance of species in riparian buffers and unharvested riparian sites.



A replicated, controlled, site comparison study in 2001 of amphibians in 41 forest streams in Washington, USA (Pollett, MacCracken & MacMahon 2010) found that where buffers were retained during clearcutting, densities of two of three species were significantly higher. Densities were significantly higher with buffers than without for tailed frogs Ascaphus truei (0.4 vs 0/m2) and cascade torrent salamander Rhyacotriton cascadae (0.5 vs 0.2). For both species, densities were significantly higher in unharvested forests (0.7 and 1.5/m2 respectively) but not secondary forests (0.2 and 0.6). In contrast, giant salamander Dicamptodon spp. densities were significantly lower in buffered (0.2/m2) than unbuffered streams and secondary forests (0.3/m2). Densities in unharvested forests (0.2) were significantly lower than the average for managed forests. Nine to 12 streams in each of four management types were sampled: clearcuts (≤10 years old) with 5–23 m wide buffers or without buffers, second-growth forest (≥35 years old) and unharvested forest. Amphibians were monitored within six 2 m long plots within 45–55m sub-sections of streams in June–August 2001.



A replicated, controlled, before-and-after study in 1992–2004 of conifer plantations in Washington, USA (Hawkes & Gregory 2012) found that retaining riparian buffers during harvest had mixed effects on amphibians. Western red-backed salamanders Plethodon vehiculum and ensatinas Ensatina eschscholtzii appeared to benefit from riparian buffers. However, coastal tailed frogs Ascaphus truei declined significantly immediately after harvest at sites with wide buffers and 10 years after treatment the species was almost locally extinct at narrow and wide buffered sites. For other species there was suggestion of treatment effects, but analyses were confounded by patterns of natural population changes. In 1992, 18 sites (33–50 ha) were selected and assigned to three treatments: forest harvested with a riparian buffer of approximately 8 m or a wider buffer (plus wildlife reserve trees/logs) and control sites of previously logged second-growth forests. Streams were 2–6 m wide and had clear-cutting of 15 ha either side. Amphibians were monitored in October–November before harvest (1992–1993), 2-years after (1995–1996) and 10-years after harvest (2003–2004). Eighteen pairs of pitfall traps were placed in buffers and adjacent habitat.


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.