Cease livestock grazing: Grassland & shrubland
Overall effectiveness category Awaiting assessment
Number of studies: 15
Background information and definitions
Grazing by livestock reduces vegetation height and ground cover, alters plant abundance and diversity, creating openings for seed growth and preventing reed or shrub growth. These changes can have beneficial (Tesauro & Ehrenfeld 2007) or detrimental effects (Howland et al. 2014) on reptile populations depending on the reptile species, grazing intensity, timing and conjunction with burning regimes. Studies included in this intervention measure the impact of ceasing grazing on reptiles. Studies that compare the effects of varying intensities of grazing or different types of grazing regimes on reptiles are included under the intervention Modify grazing regime.
For interventions that aim to reduce the detrimental effects of grazing by wild herbivores see Threat: Invasive or problematic species - Remove or control invasive or problem herbivores and seed eaters.
Howland B., Stojanovic D., Gordon I.J., Manning A.D., Fletcher D. & Lindenmayer D.B. (2014) Eaten out of house and home: impacts of grazing on ground-dwelling reptiles in Australian grasslands and grassy woodlands. PLoS One, 9, e105966.
Tesauro J. & Ehrenfeld D. (2007) The effects of livestock grazing on the bog turtle [Glyptemys (=Clemmys) muhlenbergii]. Herpetologica, 63, 293–300.
Supporting evidence from individual studies
A site comparison study (year not provided) in the Mojave Desert California, USA (Busack & Bury 1974) found that an ungrazed site had twice the number of lizards and two more species compared to a grazed site. Results were not statistically tested. In total, 36 lizards from five species were recorded in an ungrazed site (4 desert horned lizards Phrynosoma platyrhinos, 6 zebra-tailed lizards Callisaurus draconoides, 3 long-nosed leopard lizards Gambelia wislizenii, 11 common side-blotched lizards Uta stansburiana, 12 western whiptails Aspidoscelis tigris) compared to 17 lizards from three species in a grazed site (11 zebra-tailed lizards, 5 common side-blotched lizards, 1 western whiptail). Lizard surveys were carried out in May in a 100 x 100 m plot in one site with no grazing and in one site heavily grazed by sheep.Study and other actions tested
A site comparison study in 1989 in semi-arid grassland and oak savanna in Arizona, USA (Bock et al. 1990) found bunchgrass lizard Sceloporus scalaris slevini abundance was higher in ungrazed areas compared to grazed areas. In nine hours of searching ungrazed grassland, 41 lizards were observed compared to three lizards in nine hours of searching in grazed grassland. The ungrazed area (in a 3,160 ha ranch sanctuary) had not been grazed by livestock since 1967. The adjacent grazed area had been grazed for over a century. Abundance was determined by active searches counting the numbers of lizards over nine days in August 1989.Study and other actions tested
A before-and-after study in 1986–1993 on Mana Island, New Zealand (Newman 1994) found that following removal of cattle (and cessation of grazing), and subsequent eradication of an invasive mouse Mus musculus the abundance of one of four lizard species decreased, two remained stable, and one increased. Before-and-after comparisons were not statistically tested. Fewer copper skinks Cyclodina aenea were caught after grazing stopped (1–4 captures/100 trap nights) compared to before (9 captures/100 trap nights). In the four years following mouse eradication (when grazers were still absent), the number of McGregor's skinks Cyclodina macgregori increased from 1 to 10 captures/100 trap nights, though numbers were similar during grazing and in the first two years after grazing stopped (6–8 captures/100 trap nights). More common geckos Hoplodactylus maculatus were caught when there was no grazing and mice had been eradicated (35–70 captures/100 trap nights) compared to before eradication (15 captures/100 trap nights) and during grazing (5 captures/100 trap nights). A similar number of common skinks Leiolopisma nigriplantare polychrome were captured after grazing ceased (after: 6–21 captures/100 trap nights) compared to before (12 captures/100 trap nights). Cattle were removed from the island in 1986–1987, and the mouse population was eradicated using poison baits in 1989–1990. In 1985–1993, lizards were trapped annually (3–8 sessions/year; 2–4 days trapping/session) using pitfall traps (582–4,066 trap nights/session) that were deployed across 27 trapping stations around the island.Study and other actions tested
A replicated, site comparison study in 1994–1996 in desert shrub and grassland in south-central California, USA (Brooks 1999) found that lizard abundance and species richness was higher or similar inside a protected area fenced to prevent sheep grazing, compared to grazed areas outside of the fenceline, depending on survey month and site. Lizard abundance was higher in three of six survey comparisons inside a fenced protected area without sheep grazing (4–10 lizards/transect) compared to outside of it (2–4 lizards/transect) but similar in the remaining three comparisons (inside: 2–5 lizards/transect; outside: 1–3 lizards/transect; see original paper for details). Lizard species richness was higher in one of six comparisons inside the protected area (2 species/transect) compared to outside of it (1 species/transect) but similar in the remaining five comparisons (inside: 2–3 species/transect; outside: 1–3 species/transect; see original paper for details). In 1994, two sites were selected near the north-eastern and southern boundary of the Desert Tortoise Research Natural Area (where off-road vehicles were prohibited from 1973, sheep grazing prohibited from 1978 and the boundary was fenced in 1980). Two 2.25 ha plots were established/site: one ≥400m inside the boundary and one outside the boundary (used by off-road vehicles until 1980 and grazed by sheep until 1994). In each plot, lizards were surveyed using 1.25 km transects in July 1994 and May and July 1995 (six surveys/site).Study and other actions tested
A replicated, controlled study in 1994–1997 in four pastures in chaparral shrubland in Texas, USA (Kazmaier et al. 2001) found that excluding grazing from pastures resulted in similar abundance, survival and size of Texas tortoises Gopherus berlandieri compared to pastures with rotational grazing. The abundance of tortoises was similar in ungrazed (4 tortoises/100 km and 3 tortoises/10 hours) and grazed pastures (5 tortoises/100 km and 4 tortoises/10 hours). Annual survival of radio tracked individuals was similar in ungrazed (70–83%) and grazed pastures (73–84%), and size and growth were also similar (see paper for details). Two pastures each were ungrazed and grazed. Grazing was rotational (October–May) and stocking densities varied (0.2–0.6 animal units/ha; animal unit = 2 steers), though impacts on herbaceous vegetation were similar. Tortoises were counted by driving along tracks (recording the distance and time travelled) throughout the pastures between 7 April 1994 and 12 October 1997. Search effort was equal across months and time of day, and between grazed and ungrazed pastures. Forty-seven tortoises were also radiotracked.Study and other actions tested
A replicated, randomized, controlled study in 1998–1999 in streams through pasture and associated riparian areas on farms in Pennsylvania, USA (Homyack & Giuliano 2002) found that excluding livestock grazing did not increase combined reptile and amphibian species richness or abundance within 1–3 years. Overall reptile and amphibian species richness and abundance, and overall snake and turtle abundances were similar between sites fenced to exclude livestock and unfenced grazed sites (results reported as statistical tests). Of three snake species detected, abundances were higher in fenced compared to unfenced sites for northern queen Regina septemvittata (fenced: 6 individuals/site; unfenced: 2) and eastern garter snake Thamnophis sirtalis (fenced: 5; unfenced: 2) and similar for northern water snake Nerodia sipedon (fenced: 5; unfenced: 4). Ten fenced and ungrazed and 10 unfenced and grazed streams and riparian areas (100 m long, 10–15 m wide) were compared on private farms. All ungrazed sites had been grazed until they were fenced to exclude livestock 1–2 years prior to 1998 (4-strand electric fence). Unfenced stream sites and surrounding pastures were grazed continuously with an average stocking rate of 0.4 animals/ha. Reptiles were monitored using drift fences with pitfall traps set perpendicular to streams, coverboards and opportunistically using hand captures. Traps were checked 3–4 times a week from April–July 1998 and 1999.Study and other actions tested
A replicated, controlled, before-and-after study in 1994–1997 in three sites of chenopod scrubland in South Australia, Australia (Read 2002), found that overall reptile species richness and capture rates were similar at ungrazed sites compared to those under short-term intensive grazing, but that capture rates of one species increased one year after intensive grazing. Overall reptile captures and species richness were similar in an ungrazed area and in paddocks with short-term intensive grazing, both immediately before and after grazing, and one year after grazing (results reported as statistical model outputs). Central netted dragon Ctenophorus nuchalis capture rates remained similar in the ungrazed area (0.5–0.6 individuals/plot) and grazed areas immediately before and after grazing (before: 0.3–0.6 individuals/plot, after 0.5–0.5), but were lower one year later in the ungrazed area (0.4) than in the grazed area (1.1–1.5). See paper for details of other species capture rates. Reptiles were surveyed in three sites: an ungrazed area and two adjacent short-term intensively grazed paddocks (20 ha each). Intensive grazing consisted of releasing 70–80 cattle into each paddock for 6–18 days in winter and summer 1995. Reptile surveys took place twice before, twice immediately after, and twice one year after grazing using drift fences with pitfall traps open for 10 days at a time (18 fence-trap plots in grazed and in 12 in ungrazed paddocks). Captured lizards were marked with unique toe clips.Study and other actions tested
A site comparison study in 1999–2000 in grazed and cultivated semi-stable sand dunes in Zaranik Protected Area in North Sinai, Egypt (Attum et al. 2006) found that excluding livestock grazing increased Be’er Sheva fringe-fingered lizard Acanthodactylus longipes abundance. Fringe-fingered lizards were more than three times as abundant in ungrazed fenced (29 individuals/site) compared to unfenced grazed sites (9 individuals/site). Lizards spent less time moving and were observed further away from the nearest vegetation in ungrazed fenced compared to unfenced grazed sites (60 vs 38 seconds, 105 vs 55 cm). Lizards were sampled in three sites protected by fences in a protected area and three unfenced sites subject to grazing and low-impact watermelon farming. All sites were 50 m x 50 m. Sites were sampled two to four times monthly, between September 1999 and September 2000. Lizards were visually observed for three minutes and captured for measurement and marking when possible.Study and other actions tested
A paired sites, controlled study in 2002–2003 of semi-desert shrub and grassland, south-eastern Arizona, USA (Castellano & Valone 2006) found that lizard species richness was similar in ungrazed and grazed sites, but that some species abundances were higher in ungrazed sites, depending on the vegetation type. Species richness was the same in ungrazed and grazed sites (both 7–8 species). In tarbrush-dominated vegetation, two species were more abundant in ungrazed (eastern fence lizards Sceloporus undulatus: 17 individuals; common side-blotched lizards Uta stansburiana: 21) than grazed land (eastern fence: 2; side-blotched: 1), one species was less abundant (round-tailed horned lizards Phrynosoma modestum ungrazed: 3, grazed: 13) and one species had similar abundances (western whiptail Cnemidophorus tigris: 31, 37). For three species sample sizes were too small for analysis (desert spiny Sceloporus magister: 0, 1; ornate tree Urosaurus ornatus: 2, 1; grassland whiptail lizards Cnemidophorus uniparens: 7, 3). In creosote-dominated vegetation, four of eight lizard species abundances were similar in ungrazed and grazed land (eastern fence 26, 17; side-blotched 34, 29; round-tailed horned: 10, 11; western whiptail: 85, 82). For four species, sample sizes were too small for analysis (desert spiny lizard: 2, 3; ornate tree lizard: 3, 4; western banded gecko Coleonyx variegatus: 1, 1; grassland whiptail: 8, 2). A 9 ha area was fenced (post and barbed wire) to exclude livestock in 1958. Grazing continued outside of the enclosure. Lizards were monitored using pitfall traps along 12–13 transects (3–5 traps/transect) that extended from outside to inside the enclosure (60–100 m each side of the enclosure, 20–250 m apart) in August 2002 (728 trap nights) and May–August 2003 (4,620 trap nights). Transects included two vegetation types: tarbrush (1,428 trap nights) and creosote (3,920 trap nights). All lizards, other than western banded geckos, were individually marked with toe clips. Only adults were included in the analysis.Study and other actions tested
A paired sites, controlled, before-and-after study in 1993–1996 and 2007 in chenopod scrubland in South Australia, Australia (Read & Cunningham 2010) found that the effect of ceasing grazing on abundance varied depending on the species. After fencing to exclude livestock, one gecko species increased (knob-tailed gecko Nephrurus levis after fencing: 3.3 individuals/plot; before fencing: 0.3–0.5 individuals/plot) and two geckos decreased (tessellated gecko Diplodactylus tessellatus after fencing: 0.0; before fencing: 1.2–1.7; variable fat-tailed gecko Diplodactylus conspicullatus after fencing: 0.4; before fencing: 1.5–1.9) in abundance compared to beforehand when the same plots were grazed. The abundance of five other species remained similar after grazers were excluded (see paper for details). Four paired sites of differing grazing pressure were set out in 1993 (low intensity grazing: <12 cattle dung/ha; medium: 12–100; high: >120). After four years, three of the eight grazing pressure sites were fenced to exclude cattle and predators. Reptiles were sampled for 10 days in summer from 1993–1996 and 2007 using 300 mm long flymesh drift fences with 13 unbaited pitfall traps (500 mm deep x 150 mm wide, 8 m apart).Study and other actions tested
A replicated, controlled study in 1997–2006 in scrub and grassland in central California, USA (Germano et al. 2012) found the abundances of one of three reptiles increased more slowly in ungrazed plots compared to grazed plots following burning (natural and prescribed). The abundance of blunt-nosed leopard lizards Gambelia sila increased at a slower rate in ungrazed plots (1 extra individuals/year) compared to grazed plots (7 extra individuals/year). The change in abundances of two other species (western whiptail lizards Aspidoscelis tigris and side-blotched lizards Uta stansburiana), and the overall abundances of all three species did not differ between ungrazed and grazed plots (see original paper for details). Four 3 km2 areas in a single site were established and grazed from December–April in 1998–2001 and 2005–2006. Grazing intensity varied between years and the whole site had been burned (natural and prescribed fire) in 1997. Within each area, a 25 ha plot was fenced to exclude livestock (ungrazed). Day-active lizards were surveyed visually within a 9 ha grid in each grazed and ungrazed area on 10 days in May–July in 1997–2006 (800 survey days).Study and other actions tested
A replicated, randomized, site comparison study in 2011 in desert shrub and grassland in the western Mojave Desert, California, USA (Berry et al. 2014) found that Agassiz’s desert tortoises Gopherus agassizii were more abundant and had a lower mortality rate in a protected area fenced to exclude livestock grazing and recreational vehicle use. Desert tortoise densities were approximately six-times higher in the most protected area, the Tortoise Natural Area (15 tortoises/km2) than in designated tortoise critical habitat (2 tortoises/km2) and four-times higher than on private lands (4 tortoises/km2). Tortoise annual death rates over the preceding four years were estimated as lowest in the Tortoise Natural Area (3%/year) compared to private lands (6%/year) or in critical habitat (20%/year, results not statistically compared). Tortoises were surveyed in 240 1 ha plots across three different management areas (80 plots/area): Tortoise Natural Area (1973: closed to recreational vehicles; 1980: fully enclosed and closed to mining and livestock grazing, 2010: 12 km of fencing extended to prevent tortoises leaving), tortoise critical habitat areas (1994: recreational vehicle use restricted but not enforced with some annual closures, 1990: closed to sheep grazing) and private lands (unregulated sheep grazing, intensive recreational vehicle use, hunting and rubbish dumping). In April–May 2011 plots were surveyed on foot twice in a day for live or dead tortoises and field signs.Study and other actions tested
A controlled, before-and-after study in 1997–2013 in an area of mixed dry and wet heathland in Dorset, UK (Reading & Jofré 2015, same experimental set-up as Reading & Jofré 2016) found that an area where grazing cattle were excluded with a fence had more smooth snakes Coronella austriaca compared to an area where grazing continued. Over four years after grazing cattle were excluded, more smooth snakes were found in the ungrazed area (28 snakes) compared to the grazed area (16 snakes). During the previous 13 years when the whole area was grazed, the number of snakes caught in each area was similar during 12 of 13 years (3–8 snakes/year). In February 2010, a fence was erected to exclude cattle from a 6 ha area of heathland that had been grazed by cattle Bos taurus during May–September since 1997 (0.1–0.3 cows/ha). The remaining 4 ha continued to be grazed after the fence was erected. In 1997–2013, annual surveys for reptiles were conducted (21 surveys/year, though only 18 in 1997 and three in 2002) by randomly placing groups of 37 artificial refuges (corrugated steel sheets) in a hexagonal pattern (5–7 groups of refuges in the ungrazed area; four groups in the grazed area). All refuges were checked for reptiles, and smooth snakes were individually marked using PIT tags.Study and other actions tested
A controlled study in 2010–2013 in an area of mixed dry and wet heathland in Dorset, UK (Reading & Jofré 2016, same experimental set-up as Reading & Jofré 2015) found that three of four reptile species were more abundant in ungrazed compared to grazed areas, and the fourth species occurred at similar numbers in both areas. The ungrazed area contained more grass snakes Natrix natrix (2/plot), slow worms Anguis fragilis (67/plot) and common lizards Zootoca vivipara (13/plot) than the grazed area (grass snakes: 1/plot; slow worms: 29/plot; common lizards: 6/plot), whereas a similar number of sand lizards Lacerta agilis were found in the ungrazed (3/plot) and grazed (6/plot) areas. In February 2010, a fence was erected to exclude cattle from a 6 ha area of heathland that had been grazed by cattle Bos taurus. The remaining 4 ha continued to be grazed after the fence was erected. In 2010–2013, annual surveys for reptiles were conducted (21 surveys/year) by randomly placing 11 groups of 37 artificial refuges (407 refuges in total) during April–October (seven groups of refuges in the ungrazed area; four groups in the grazed area). The number of reptiles of each species was recorded at each visit.Study and other actions tested
A replicated, site-comparison study in 1997–2007 in shrub and woodland in south eastern Australia, Australia (Haby & Brandle 2018) found that ungrazed and grazed sites had similar combined reptile and small mammal species richness. Over 11 years, reptile and small mammal species richness remained similar in ungrazed (0.03 species/100 trap nights/year) and grazed shrubland (0.04 species/100 trap nights/year). Over the same time period, livestock removal did not affect the change in overall reptile and small mammal abundance over time in shrubland (no livestock: 0.02 individuals/100 trap nights/year; with livestock: 0.11) In 1997–2007, reptiles and small mammals were surveyed in two shrubland sites (degraded chenopod shrubland dominated by A. victoriae.) with historical but no current domestic livestock grazing and two sites with livestock (sheep and/or cattle) grazing. Reptiles were surveyed using pitfall traps one–three times/year (22 surveys).Study and other actions tested