Use prescribed burning to control grass
Overall effectiveness category Unknown effectiveness (limited evidence)
Number of studies: 2
Background information and definitions
This section considers the use of prescribed fire to control problematic grass species in shrublands. Fire may damage or kill grasses therefore reducing competition between with other shrubland vegetation. This may allow shrubland plants to increase in abundance.
Supporting evidence from individual studies
A randomized, replicated, controlled, paired, before-and-after study in 1996-1999 in wet heathland in Northumberland, UK (Ross et al. 2003) found that burning to reduce the cover of purple-moor grass Molinia caerulea reduced cover of common heather Calluna vulgaris but did not reduce cover of purple-moor grass. Cover of common heather was lower after burning than before burning (after: 20-50%, before: 71-97%) and was lower than in plots that were not burned (68-93%). Cover of purple-moor grass in plots that had been burned was not significantly different to that in plots that had not been burned (no data presented). In April 1996 six 10 x 10 m plots were burned and six plots were not burned. In 995 five 1 m2 quadrats were established in each plot and vegetation was surveyed in July 1995 and 1997-1999. The area was used for livestock and density of sheep varied from 0.66 to 1.5 ewes/ha.Study and other actions tested
A randomized, replicated, controlled study in 1995–2000 in four moorland sites in the UK (Marrs et al. 2004) found that prescribed burning to reduce the cover of purple moor grass Molinia caerulea initially reduced vegetation height, but this subsequently recovered. Immediately after prescribed burning vegetation height was lower in burned plots than in unburned plots, however after five years there was no longer a difference in height between burned and unburned plots (no data reported). In 1995 two blocks were established on each site and half of each block was burned. Fencing was established to limit grazing in two plots within each burned area but one plot was left unfenced. Within each plot the herbicide glyphosate was applied in two subplots and one subplot was not sprayed. Vegetation height was measured in 20 random locations with a sward stick in each plot.Study and other actions tested