RESTORATION EVIDENCE

Action: Allow shrubland to regenerate without active management Shrubland and Heathland Conservation

Key messages

Read our guidance on Key messages before continuing

  • Five before-and-after trials (two of which were replicated) in the USA, UK, and Norway, found that allowing shrubland to recover after fire without any active management increased shrub cover or biomass. One replicated, paired, site comparison in the USA found that sites that were allowed to recover without active restoration had similar shrub cover to unburned areas. One controlled, before-and-after trial in the USA found no increase in shrub cover. One before-and-after trial in Norway found an increase in heather height. One before-and-after trial in Spain found that there was an increase in seedlings for one of three shrub species. Two replicated, randomized, controlled, before-and-after trials in Spain and Portugal found that there was an increase in the cover of woody plant species. One before-and-after study in Spain found that cover of woody plants increased, but the number of woody plant species did not. One replicated, before-and-after study in South Africa found that the height of three protea species increased after recovery from fire.
  • One before-and-after trial in South Africa found that there was an increase in vegetation cover, but not in the number of plant species. One before-and-after trial in South Africa found an increase in a minority of plant species. 
  • Two before-and-after trials in the USA and UK found that allowing shrubland to recover after fire without active management resulted in a decrease in grass cover or biomass. One controlled, before-and-after trial in the USA found an increase in the cover of a minority of grass species. One before-and-after study in Spain found that cover of herbaceous species declined. One replicated, before-and-after study in the UK found mixed effects on cover of wavy hair grass.
  • One controlled, before-and-after trial in the USA found no increase in forb cover. One replicated, randomized, controlled before-and-after trial in Spain found that herb cover declined after allowing recovery of shrubland after fire.

Supporting evidence from individual studies

1 

A before-and-after trial in 1936–1966 in a sagebrush scrub shrubland affected by fire in Idaho, USA (Harniss & Murray 1973) found that allowing shrubland to recover from disturbance without any active restoration increased biomass of big sagebrush Artemisia tridentata and decreased grass biomass after 30 years. After 30 years of recovery, biomass of big sagebrush was higher (325 kg/ha) than in the same areas immediately after fire had occurred (2 kg/ha). However, the areas’ grass biomass did not differ significantly after 30 years of recovery (164 kg/ha) from grass biomass in the same areas immediately following fire (158 kg/ha). In 1936 the shrublands were burned. Following this vegetation cover was assessed in four hundred 9.3 m2 quadrats in 1936, 1937, 1939, 1948, and 1966.

2 

A before-and-after trial in 1961–1971 in a heathland affected by fire in Dorset, UK (Brian et al. 1976) found that allowing shrubland to recover without active restoration increased cover of two of four shrub species, and decreased cover of two of two grass species and bracken Pteridium aquilinum. After the heathland was allowed to recover for 10 years, cover of common heather Calluna vulgaris and bell heather Erica cinerea increased, but cover of cross-leaved heath Erica tetralix and dwarf gorse Ulex minor did not increase (data presented as index). After a 10 years recovery, the cover of purple-moor grass Molinia caerulea and bristle bent Agrostis setacea was not significantly different to cover immediately after the fire (data presented as index). However, the cover of bracken was lower than immediately after fire than after 10 years of recovery (data presented as index). In 1961 and 1971 vegetation cover of plant species was estimated at 157 points.         

3 

A before-and-after trial in 1980–1981 in grassy fynbos habitat previously affected by fire in Eastern Cape Province, South Africa (Richardson et al. 1984) found that allowing recovery of vegetation without any active restoration increased the abundance of three of twenty plant species. After one year, three of twenty plant species had higher abundance (3–12 plants/m2) than immediately after the fire (2–6 plants/m2). In August 1980 fire burned the majority of the site. Twelve 1 m2 quadrats were placed at the site and abundance of plants was recorded every month between December 1980 and December 1981.

4 

A controlled, before-and-after trial in 1981–1983 in a sagebrush scrub habitat that had been burnt in a wildfire in Utah, USA (West & Hassan 1985) found that allowing shrubland to recover without active restoration did not increase cover of shrub, or forb species, but did increase cover of two of eight grass species. After two years of recovery, the cover of five of five shrubs (0–1%) was not significantly different than immediately after burning (0%), and was lower than in unburned areas for one of five shrubs (3%). The same pattern was true for seven of seven forb species (after recovery: 0%, immediately after fire: 0%). Cover of two of eight grass species was higher after two years of recovery (5–50%) than immediately after fire (0%). Wildfire occurred in July 1981. Cover of plants was assessed visually in four 1000 m2 plots every year in 1981–1983.

5 

A before-and-after study in 1989-1994 in rockrose shrublands in northern Spain (Tarrega et al. 1995) found that allowing shrubland to recover after severe disturbance without any active restoration increased the cover of woody plants, reduced the number of species and cover of herbaceous plants, but did not increase the number of woody plant species after five years. In three of three cases, after five years, cover of woody plants in plots was higher (50-64% cover) than the in same plots immediately after they were disturbed (1-25% cover), but was lower than the same plots before they were disturbed (79-91% cover). In three of three cases the number and cover of herbaceous species after five years was lower (species: 2-7; cover: 2-12%) than that immediately after disturbance (species: 6-16; cover: 17-61%), but cover after five years was higher than herbaceous cover prior to disturbance (0-3% cover). However, in three of three cases the number of woody species did not differ significantly after five years (2-3 species) and the same areas immediately after disturbance (1-3 species). Three 100 m2 plots were established in the shrublands in 1989 and were either burnt, ploughed, or cut. Each year in 1989-1994 five 1 m2 quadrats were placed in each plot and vegetation estimated visually.

6 

A before-and-after trial in 1991–1994 in a heathland that had previously been burnt by wildfire in southern Spain (Ojeda et al. 1996) found that allowing recovery without active restoration increased the number of seedlings of common heather Calluna vulgaris, but not of rock rose Cistus populifolius and Genista triacanthos. After three years of recovery, the number of seedlings of heather (29 seedlings/m2) was higher than found in the first year after fire (0 seedlings/m2). However, the numbers of rock rose (7 seedlings/m2), and Genista triacanthos (7 seedlings/m2) seedlings was not higher after three years than in the first year after fire (rock rose: 8 seedlings/m2, Genista triacanthos: 6 seedlings/m2). A wildfire occurred in 1991. The abundance of seedlings was recorded annually in 50 randomly located 1 m2 quadrats in 1992–1994.

7 

A replicated, before-and-after trial in 1996–1998 in two heathlands in the UK (Britton et al. 2000) found that allowing shrubland to recover from disturbance without any active restoration increased cover of heather Calluna vulgaris in two of two sites, but increased cover of wavy-hair grass Deschampsia flexuosa in one of two sites. After two years of recovery, heather cover increased from 0% to 21–85% in both sites. However, at one site, the cover of wavy-hair grass increased from 0% to 8%, while at the other site wavy-hair grass cover was 0% before recovery and 0% after two years of recovery. All vegetation in randomly placed plots of 1 m2, 0.25 m2, and 0.0625 m2 was removed and the top 15 cm of soil broken up with a spade. Vegetation cover in each plot was recorded in April and October of each year in 1996–1998.

8 

A replicated, before-and-after trial in 1993–1998 in 90 sage scrub and chaparral shrubland sites that had been burned by wildfire in California, USA (Keeley et al. 2005) found that allowing shrublands to recover without active restoration increased shrub cover. After five years of recovery, shrub cover was higher (45–92%) than it had been immediately after wildfires (4–15%). Sites were 0.1 ha areas where wildfires had burned in the previous year. At each site twenty 1 m2 quadrats were used to estimate plant cover and density.

9 

A before-and-after trial in 1999–2002 in 16 heathlands that had previously been burned in central Norway (Nilsen et al. 2005) found that allowing recovery without any active restoration increased the cover and height of common heather Calluna vulgaris. Areas that had been allowed to recover had higher cover one year after burning than immediately after burning (data presented as log units). Common heather plants were taller after one year of recovery (1.8–6.8 cm) than immediately after burning (0.2–4.2 cm). Sixteen heathland sites were burned in 1999–2001. Three 1 m2 plots were placed at each site and vegetation cover and height were monitored in 2001–2002.

10 

A before-and-after trial in 2000–2004 in two fynbos habitats previously affected by fire in Western Cape, South Africa (Reinecke et al. 2008) found that allowing shrubland to recover without active restoration increased vegetation cover but not plant species richness. After four years of recovery, vegetation cover was higher (84–94 cm2) than immediately after the site was burned (62–27 cm2). However, after four years of recovery plant species richness did not increase significantly (before: 3–5 species, after: 3–4 species). In 2000 wildfires burned both sites. Two 1 m wide transects consisting of forty-nine to seventy-eight 1 m2 quadrats were used to survey vegetation cover in 2001, 2002, and 2004.

11 

A replicated, randomized, controlled study between 1998 and 2008 in sagebrush scrub in Oregon, USA (Boyd & Svejcar 2011) found that allowing shrubland to recover from disturbance without any active restoration increased shrub cover to a level similar to that found in undisturbed shrublands, while annual grass cover was higher than seen in undisturbed shrublands and perennial grass cover was lower. Ten years after all vegetation was removed shrub cover had increased from 2% to 22%, similar to the cover of 24% seen in undisturbed shrublands. Over 10 years perennial grass cover increased from 0% to 10% but was still lower than the cover of 25% found in undisturbed shrubland. Annual grass cover increased from 2% to 7% and this was higher than the 2% cover seen in undisturbed shrubland. However, annual forb cover declined from 20% to 3%, similar to the cover of 1% seen in undisturbed shrubland. Perennial forb cover did not show a significant change over time (1% vs 2%) and was not significantly different from cover seen in undisturbed shrubland (1%). In 1998 all vegetation was removed from five 36 m2 plots using glyphosate herbicide with a further five plots receiving no treatment. Vegetation cover was estimated in five 0.2 m2 quadrats in each experimental plot between 1999 and 2008.

12 

A replicated, before-and-after study in 1996–2011 in two fynbos sites in South Africa (Kraaij et al. 2013) that had previously been burned found that allowing the sites to recover without any active management led to an increase in the height of three protea species. At one site Leucadendron eucalyptifolium and forest sugarbush Protea mundii were taller six years after burning (L. eucalyptifolium: 140 cm, forest sugarbush: 127 cm) than one year after burning (L. eucalyptifolium: 62 cm, forest sugarbush: 62 cm). At the other site oleanderleaf protea Protea neriifolia were taller after 15 years (148 cm) than 10 years after burning (92 cm). One of the sites was burned in a prescribed fire and the other in a wildfire. The height of 100 seedlings of oleanderleaf protea was recorded in April 1996 at one site and the height of 100 L. eucalyptifolium and forest sugarbush seedlings was recorded in 2008 at the other. Plant height was recorded annually until 2011.

13 

A replicated, randomized, controlled, before-and-after trial in a shrubland in southern Spain (Cespedes et al. 2014) found that allowing shrubland to recover from fire without any active restoration increased cover of woody plants but reduced cover of herb species after three years. Three years after fire, cover of woody species was higher (67%) than one year after fire (21%) and was not significantly different from cover before fire occurred (75%) or cover in unburnt plots (73%). The cover of herb species was lower three years after fire (18%) than one year after fire (54%), but cover was not significantly different to unburnt plots (14%). Twelve 625 m2 plots were established and randomly assigned to be burned, or left unburnt. Before burning and for three years afterwards twenty 1 m2 quadrats per plot were used to assess cover of plant species. Seedling density was assessed using 25 cm x 25 cm quadrats.

14 

A replicated, paired, site comparison study of seven shrublands that had been previously burned in New Mexico and Texas, USA (Ahlstrand 1982) found that sites that were allowed to recover without active restoration had a similar shrub cover to unburned sites, but higher grass cover. Shrub cover did not differ between sites after seven years (burned: 7%; unburned: 9%). After seven years of recovery grass cover was higher (53%) than in unburned sites (35%). Seven sites that had been burned between three and seven years ago and seven paired unburned sites were selected. Plant species and their percentage cover were recorded within fifty 0.1 m2 plots/site.

15 

A replicated, randomized, controlled, before-and-after trial between 2002 and 2006 in a shrubland in Central Portugal (Cespedes et al. 2014) found that allowing shrubland to recover from fire without any active restoration increased the cover of woody species after four years but not to pre-fire levels. Four years after fire, cover of woody plant species was higher (70%) than one year after fire (47%) but not as high as before fire occurred (98%). The cover of perennial herb species was higher four years after fire (9%) than one year after fire (5%), both of which were higher than pre-fire levels (1%). Four years after fire the cover of bracken Pteridium aquilinum was higher (2%) than pre-fire levels (0%). Twelve 0.2 hectare plots were established in the shrubland, eight of which were subsequently burned and four of which were left unburned. Vegetation cover was estimated using twenty 1 m2 quadrats which were surveyed before burning and annually for four years afterwards.

Referenced papers

Please cite as:

Martin P.A., Rocha R., Smith R.K. & Sutherland W.J. (2018) Shrubland and Heathland Conservation. Pages 447-494 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.