Use prescribed burning on savannas
Overall effectiveness category Trade-off between benefit and harms
Number of studies: 5
Background information and definitions
Periodic wildfires and grazing/browsing by wild animals maintained open savanna ecosystems in parts of the world, retarding woody plant encroachment and succession to woodland. In some regions where these natural processes have been lost, and where active fire suppression may have also occurred, rotational burning may be used to reinstate and maintain more open habitat conditions amenable. Other methods commonly used include thinning of trees and shrubs, timber harvesting, and livestock grazing.
There is a continuum between habitats classified as ‘savannas’ and those classified as ‘forests’ and ‘grasslands’. Relevant information for management of habitats with both trees and grasses may, therefore, be found in all three sections.
Supporting evidence from individual studies
A small study in Minnesota, USA (Cox 1987), found that eastern bluebird Sialia sialis clutches in two nest boxes in an area that underwent prescribed burning both had 100% success (i.e. all eggs produced fledglings), compared with an average success of 93% for 23 nest boxes nearby. Flame height was 1 m or less and did not reach the boxes at 1.2 and 1.4 m above ground. Adult bluebirds left their boxes as the fire approached, returning when it had passed.Study and other actions tested
At Cedar Creek Natural History Area, Minnesota, USA (Davis et al. 2000), a replicated controlled study of oak savanna restoration by prescribed burning (initiated in 1964) found that ‘open country’ bird abundance increased as restoration progressed. Seven units (8-18 ha) were subject to one of seven burn frequencies, ranging from nearly every year to no burning over the previous 31 years. Burns were conducted in spring (except two in late summer). Bird species richness in the two unburned units (17 and 23 species) was lower than that of two frequently burned units (30 and 32) in June 1995 and similar in 1996. As woodland became more open, upper tree canopy insectivores declined, whilst omnivorous birds, particularly ground and lower canopy foragers increased. Woodpeckers increased as standing dead tree abundance increased.Study and other actions tested
A replicated, paired study in oak Quercus spp. savanna in Illinois, USA (Brawn 2006), found that bird community composition was significantly different in areas with prescribed burning compared to closed-canopy oak forest. Of the 31 bird species analysed, 12 were more common in burned savanna and five more common in unburned forest. Twelve savanna sites maintained by burns (spring, autumn or both, on a 3-5 year rotation, with periodic removal of Acer spp. and European buckthorn Rhamnus catharticus) were paired with 12 forest sites (no burning for over 50 years). Point counts were conducted for 3-5 years (between 25 May-10 July 1995-1999) to assess bird abundance. There was no effect of burning on brood parasitism by brown-headed cowbird Molothrus ater.Study and other actions tested
A replicated, controlled study in January-March 2001 and 2004 in open eucalypt and riparian woodland along three creeks in Queensland, Australia (Valentine et al. 2007), found that plots burned in the dry season of 2000 had significantly fewer birds in 2004 than control (unburned) sites, whilst sites burned in the wet season of 2000 had higher abundances. Species richness did not vary in 2004. However, several species showed short-term changes in abundance after fire: three species were more abundant in dry-season burned sites in 2001; two were less abundant after burns. Pied butcherbirds Cracticus nigrogularis were more abundant in burned eucalypt sites and little friarbirds Philemon citreogularis were more abundant in dry season burnt sites and riparian habitat of wet season burnt sites.Study and other actions tested
In Laikipia District, Kenya, a replicated controlled study in 2005-2007 (Gregory et al. 2010) found that five burned plots of savanna had, on average, but not consistently, higher densities of birds and more species than five grazed or four unmanaged control areas (3-17 birds and 3-8 species/100 m2 for burned areas vs. 4-6 birds and 2.5-4.0 species for controls; 5-8 birds and 4-5.5 species for grazed areas). The authors note that there were no differences between treatments in drought years, and that the yearly variation in burned plots was greater than in grazed plots, suggesting that grazing may have longer term benefits. In addition, some species were only recorded in unmanaged areas. The impact of burning appeared to decrease over time.Study and other actions tested