Manage vegetation by hand (selective weeding)
Overall effectiveness category Awaiting assessment
Number of studies: 4
Background information and definitions
Many habitats depend on disturbances such as grazing or fire to reduce succession that leads to increases in woody plant species and conversion to forest. Manually removing vegetation may prevent succession and reduce shading.
Here we present studies that remove both native and non-native species.
For studies discussing the effects of grazing as part of habitat restoration, see Manage vegetation using livestock grazing. For studies discussing the effects of managing vegetation by cutting or using herbicides, see Manage vegetation by cutting or mowing and Manage vegetation using herbicide.
Supporting evidence from individual studies
A replicated, controlled, before-and-after study in 1993–1997 in shoreline habitat on a lake in Kwazulu-Natal, South Africa (Leslie & Spotila 2001) found that removing the invasive plant Chromolaena odorata from nesting sites by hand increased Nile crocodile Crocodylus niloticus successful nesting attempts over three breeding seasons. Results were not statistically tested. Known nesting sites where invasive vegetation was removed had 40% (2 out of 5 sites nested), 80% and 60% success over three breeding seasons following removal, compared to 40% nesting success before removal. Newly created nesting sites, where invasive vegetation was completely removed, had 33% (2 out of 6 sites nested), 33% and 67% success over three breeding seasons following removal, compared to 0% success before removal. Nesting success in sites where invasive vegetation was not removed was 100% (5 out of 5 sites nested), 60%, 40% and 40% over four breeding seasons. In 1993, sixteen nest sites were chosen: five known nesting sites where the invasive plant was present and manually removed from 1994; six sites newly created by manually removing all invasive vegetation and root stock (4 x 4 m area); and five where the invasive plant was present and was not removed. In 1994–1997 (three breeding seasons) invasive vegetation clearing was carried out each season. In 1993–1997, all sites were monitored using foot, boat and aerial surveys in mid-December to determine use of nesting sites.Study and other actions tested
A replicated, paired sites, controlled, before-and-after study in 2002–2007 in a site of dunes and desert scrub in California, USA (Barrows et al. 2009) found that manual removal of invasive non-native Sahara mustard Brassica tournefortii resulted in an increase in Coachella Valley fringe-toed lizard Uma inornata abundance compared to no weeding in one of three years in one of two habitat types, but flat-tailed horned lizard Phrynosoma mcallii abundance remained similar in all comparisons. In yearly comparisons, fringe-toed lizard abundance was higher in weeded plots in one of three years during or after weeding in active dunes (second year of weeding: 7 lizards/plot; not weeded: 4 lizards/plot; first & third years: 2–4 lizards/plot), but not in stabilized sand fields (weeded: 2 lizards/dune; not weeded: 1–3 lizards/dune). Overall abundance of fringe-toed lizards was higher in weeded plots (3 lizards/plot) compared to plots with no weeding (2 lizards/plot), but flat-tailed horned lizard abundance was similar in both (weeded: 0.1 lizards/plot; not weeded: 0.1 lizards/plot). Paired plots (10 x 100 m plots) of mustard removal and no mustard removal were established in stabilised sand fields (15 removal plots, 15 no removal plots) and active dunes (6 removal plots, 6 no removal plots). Mustard removal was carried out by hand in 2005–2006. Reptiles were surveyed at each site six times/year from May to July 2002–2007 in the morning using sightings and tracks left in the sand.Study and other actions tested
A randomized study in 2006–2008 in wetlands in New York, USA (Dowling et al. 2010) found that female Blanding’s turtle Emydoidea blandingii used weeded plots less frequently than tilled plots for nesting. Overall, fewer turtles nested in weeded plots (1 turtle in 2008) than in tilled plots (7 turtles in 2006; 5 turtles in 2008). In 2006, nine of 10 monitored female turtles nested, and in 2008, six turtles nested. Two turtles nested in the same physical plot each year, in spite of a change in management. In 2006, thread trailing revealed that all female turtles explored or had been placed on each plot type before choosing where to nest. Eight sites around the edge of a fenced 12 ha wetland were monitored for turtle nesting activity. Two plots (5 x 7 m each) were established at each site, and one/site was either hand weeded (90% of vegetation removed) or tilled to a depth of 15 cm (treatment randomly applied in 2006 and 2008). Nesting activity was monitored by visual searches and radio tracking or by attaching a bobbin and thread to female turtles in May and June 2006 and 2008 (10 turtles monitored in total).Study and other actions tested
A replicated, randomized, controlled study in 2007 on coastal sand dunes in St Croix, US Virgin Islands (Conrad et al. 2011) found that manually removing the native plant ‘beach morning glory’ Ipomoea pes-caprae did not increase leatherback turtle Dermochelys coriacea nest productivity compared to plots with unmanaged vegetation, but fewer hatchlings became entangled in plant roots in removal plots compared to in untreated, vegetated plots. Manually-removed vegetation plots had similar hatching success (24% hatched/total yolked eggs) and emergence success (emergence success: 20% hatched/total yolked eggs) compared to unmanaged vegetated plots (hatching success: 20%; emergence success: 14% hatched/total yolked eggs), but lower hatching and emergence success compared to naturally non-vegetated plots (hatching success: 50%; emergence success: 38% hatched/total yolked eggs). However, the number of hatchlings that became entangled in plant roots was lower in plots with vegetation removed (19 of 397, 5% hatchlings trapped) compared to unmanaged vegetated plots (36 of 314, 11% hatchlings trapped). Vegetation removal was carried out using machetes and weeding by hand until no vines remained above ground in April before the start of the nesting season. Ten manually-removed and 10 unmanaged vegetation plots (5 x 5 m) were randomly allocated across two experimental blocks. Ten unvegetated plots were established seaward of the experimental blocks. In April–May 2007, nests laid in areas of the beach liable to flooding were relocated to either plots with manually-removed vegetation, untreated vegetated plots or naturally unvegetated plots (2 nests/plot; 16–20 nests/treatment). Nylon mesh nets were placed over nests before hatching. All nests were excavated 1–3 days after the main period of hatchling emergence and the number of hatched and unhatched eggs was counted.Study and other actions tested