Natural Environment Research Council
NERC Banner

Allow natural regeneration of ground cover beneath perennial crops

Key messages

Natural enemies on crop trees and vines: Five studies (including one replicated, randomised, controlled test) from Australia, China, Italy and Portugal compared natural and bare ground covers by measuring numbers of natural enemies in fruit tree or vine canopies. Three found effects varied between groups of natural enemies, two found no difference. Two studies from Australia and France compared natural to sown ground cover and found no effect on enemies in crop canopies.
Natural enemies on the ground: Five studies (including three replicated, randomised, controlled trials) from Australia, Canada, China, France, and Spain compared natural and bare ground covers by measuring natural enemies on the ground. Two studies found more natural enemies in natural ground cover, but in one the effects were only short-term for most natural enemy groups. Three studies found mixed effects, with higher numbers of some natural enemy groups but not others. Two studies compared natural and sown ground covers, one study found more natural enemies and one found no effect.
Pests and crop damage: Four studies (three controlled, one also replicated and randomised) from Italy, Australia and China measured pests and crop damage in regenerated and bare ground covers. Two studies found fewer pests, whilst two studies found effects on pests and crop damage varied for different pest or disease groups. One study found more pests in natural than in sown ground covers.

Crops studied were apple, grape, lemon, olive and pear.

Background information and definitions

This includes studies allowing the natural regeneration of weeds beneath perennial crops to enhance natural enemy populations. This includes studies testing the impact of tillage versus no tillage (or other types of soil disturbance) and herbicide versus no herbicide under perennial crops, where these practices are used to control weeds. Studies using naturally regenerated ground cover as a control treatment to compare with other interventions (e.g. 'Grow plants that provide nectar or pollen resources' and 'Grow plants that provide supplementary prey for natural enemies') are not included here. Ground-dwelling invertebrates are frequently surveyed using pitfall traps – small pots buried in the ground up to their rim and left empty or filled with liquid preservatives or water.

Supporting evidence


A randomised, replicated, controlled, before-and-after trial in 1975-1977 in apple Malus domestica orchards in Ontario, Canada (Holliday & Hagley 1984) found more ground beetles (Carabidae) in naturally regenerated ground cover (29-98 ground beetles/m²) than in bare ground cover (4-7 ground beetles) in August 1976 to August 1977, in shallow 5 cm-deep soil samples. Soil cores to 30 cm-depth found a similar effect, with 42-305 ground beetles/m² in natural ground cover compared with 0-66 ground beetles in bare ground. More ground beetles also occurred in natural ground cover than in creeping red fescue Festuca rubra (7-54 and 28-122 ground beetles/m², in shallow and deep samples respectively) or perennial ryegrass Lolium perenne (0-34 and 28-122 ground beetles/m²) sown ground covers. Pitfall trapping found no effect of ground cover treatments. Naturally regenerated, bare (tilled), creeping red fescue and perennial ryegrass ground cover treatments were tested in 24 x 18 m plots (containing 18 apple trees), replicated five times each. Natural plots were untilled from 1975 to 1977. Other treatments were tilled weekly in May 1976 with fescue and ryegrass plots sown in early June 1976. Shallow and deep soil samples were taken at three and five locations/plot, respectively.


A controlled study in a pear Pyrus communis orchard in Drôme, France (Rieux et al. 1999) found that the ratio of beneficial to plant-eating invertebrates was similar in pear tree canopies over naturally regenerated (0.05 natural enemies to each pest), bare (0.04 natural enemies) and sown (0.06 natural enemies) ground covers. Flies (Empididae), leaf bugs and plant bugs (Miridae) were the most numerous enemies in trees over regenerated ground covers. Similar numbers of natural enemies were found in regenerated and sown plants on the ground. However, regenerated plants had nearly two times more plant-eating invertebrates than sown plants so there were 0.06 natural enemies to each pest in the former vs. 0.09 natural enemies to each pest in the latter. More natural enemies (for each pest) occurred in regenerated plants than in the tree canopies. Natural ground cover (established for 10 years), bare ground (created with glyphosate in March 1994) and sown ground cover treatments (planted in September 1993) each occupied one-third of the orchard (five rows between trees). Sown ground covers comprised ryegrass Lolium perenne, white mustard Sinapis alba and white clover Trifolium repens. Insects were sampled by beating branches in trees and using a sweepnet in ground covers.


A replicated, randomised and controlled study in an apple Malus domestica orchard in Asturies, Spain in 2000 (Minarro & Dapena 2003) found that ground beetle (Carabidae) abundance was similar in plots with naturally regenerated ground cover (33 captures/trap) and rotovated control plots (48 captures). The second most common ground beetle species, Pseudophonus rufipes, was significantly less numerous in the vegetated plots (0.13 captures/trap) than in the control (28 captures). Treatment blocks (comprising a row of 11 trees) were replicated four times in the 5 ha orchard. The regenerated ground cover was mowed three times in April-July and the control was rotovated in early spring and late August. Ground beetles were captured in pitfall traps of 6.5 cm diameter in August-November with two traps per plot.


A controlled study in 2002-2004 in a vineyard in Sardinia, Italy (Serra et al. 2006) found similar numbers of lacewings (Neuroptera) in a plot with regenerated ground cover (5-38 adults/plot) and a tilled plot (1-27 adults). Fewer spiders (Araneae) occurred in grape Vitis vinifera bunches in the ground cover plot (19-85 individuals/100 bunches) than in the tilled plot (29-108 individuals), but the ground cover plot had more spiders on vine trunks (371-440 versus 117-338 individuals/plot). Vine mealybug Planococcus ficus abundance did not differ consistently between plots, but infestations were higher in the ground cover (28-32% of bunches infested) than the tilled plot (17-18%) in 2003-2004, and crop damage was also higher in the ground cover plot in 2004 (28% versus 12% of bunches damaged). Infestations by second and third generation European grapevine moth Lobesia botrana were smaller in the ground cover plot (12-21% of bunches infested) compared to the tilled plot (20-40%) in 2002-2004. Grey mould Botrytis cinerea and sour rot Geotrichum candidum damage was also less in the ground cover (2-13% of bunches damaged) than the tilled plot (12-42%) in 2002-2003 but not in 2004. Ground cover was naturally regenerated and mowed in one 0.5 ha plot, and ploughed and grubbed to control weeds in another 0.5 ha plot.


A site comparison in 2005 and 2006 on two olive Olea europaea orchard plots coppiced in 1956 in the Bouches-du-Rhône, France (Ricard et al. 2007) found the plot with undisturbed ground cover had more spiders (Araneae) and ground beetles (Carabidae) (885 spiders, 69-206 ground beetles) than the plot where ground cover was ploughed (515 spiders, 27-53 ground beetles). There was a higher proportion of known predatory rove beetles (Staphylinidae) and ground beetles in the plot with undisturbed ground cover (19% and 30%) than in the ploughed plot (9% and 17%). Ground beetle species richness was higher in the orchard with undisturbed ground cover but the number of spider families was similar (undisturbed ground cover: 16 ground beetle species, 18 spider families; ploughed cover: 11 ground beetle species, 17 spider families). In 2005, rove beetle abundance was similar between orchards but species richness was higher in the disturbed ground cover orchard (undisturbed ground cover: 23 species; ploughed cover: 29 species). In spring 2006, rove beetle abundance was higher in the disturbed ground cover plot. One plot had permanent vegetation cover between rows and chemical weeding within the rows. Vegetation between rows in the second orchard was disturbed using a disc plough.


A replicated, randomised and controlled study in a grape Vitis vinifera vineyard in Victoria, Australia in 2003-2004 (Sharley et al. 2008) found that predatory ants (Formicidae) were more abundant in untilled plots containing resident vegetation (averaging 57 captures/plot) than tilled plots (35 captures) in the four months following tillage. Numbers of other ground-living natural enemies, including earwigs (Dermaptera), centipedes (Lithobiida), millipedes (Julida) and spiders (Araneae), were also greater in untilled than tilled plots in the first or second month after tillage, but similar thereafter. Pest antlike flower beetles (Anthicidae) were less abundant in untilled (averaging 0.6 captures/plot) than tilled (2.2 captures) plots across all months. In the canopy, parasitoid wasps (Trichogrammatidae) were more abundant in untilled (averaging 5 captures) than tilled plots (2 captures) in one month, but were similar a month later. In each of five 288 m² plots, half the area was tilled (15 cm depth) and half was left with natural resident vegetation (grasses and weeds).


A replicated, randomised, controlled study in south-eastern Australia (Bone et al. 2009) found that numbers of parasitoid (Hymenoptera), lacewing (Chrysopidae and Hemerobiidae) and ladybird (Coccinellidae) natural enemies  were similar in apple tree Malus domestica canopies over naturally regenerated and commercial grass ground covers. Damage caused by the majority of pests and diseases was similar between treatments (including apple dimpling bug and russett) but damage by Helicoverpa was significantly less for apples with naturally regenerated ground cover (causing 1% of the damage to apples) than with grass mix (5%) at one site. Apple diameter and weight were similar (73-75 mm diameter, 165-188 g) for apple trees in both treatments. Treatments were applied at three sites (two included the regeneration treatment but all three received the grass mix) and in plots of 265-288 m² replicated four times. Naturally regenerating species included a mix of flowering plants and grasses.


A replicated, randomised and controlled study in three lemon Citrus limon orchards in Oeste, Portugal in 2002-2003 (Silva et al. 2010) found that spiders (Araneae), ladybirds (Coccinellidae) and parasitoid wasps (Hymenoptera) were more abundant in lemon trees above naturally regenerated vegetation than above bare ground controls, when sampled by both beating and suction. Lacewings (Chrysopidae) were more abundant in lemon trees over naturally regenerated ground covers (3.0 individuals/25 trees) than controls (0.5 individuals) in suction samples, but beating samples found no difference. Ground cover treatments provided the highest numbers of lacewings, ladybirds and parasitoid wasps (relative to controls) in spring and summer, but not in winter. The three orchards were split into plots of 0.6 ha which were allowed to naturally regenerate or kept bare with herbicide. Regenerated plots were mown twice per year and comprised a mixture of grasses and flowering plants, dominated by annual meadow grass Poa annua.


A site comparison study in two pear Pyrus spp. orchards in Daxing District, China (Beizhou et al. 2011) found that total numbers of natural enemies were similar between plots of naturally regenerated ground cover (averaging 337 individuals/year) and bare, tilled plots (306 individuals/year) during March to September, 2006-2008. Pest numbers were lower in regenerated (averaging 2113 individuals/year) than bare, tilled plots (3214 individuals/year). The seven-spot ladybird Coccinella septempunctata, predatory mite Phytoseiulus persimilis and green lacewing Chrysoperla sinica were the dominant natural enemies on ground cover plants and were more abundant in regenerated than tilled plots from around early June to mid-July. Three plots of 50 x 67 m in one orchard were allowed to grow natural grasses (Poaceae) and were compared with three plots of tilled bare ground in a separate orchard. Invertebrates were counted using visual surveys, canopy traps and sweeps of ground cover vegetation.

Referenced papers

Please cite as:

Wright, H.L., Ashpole, J.E., Dicks, L.V., Hutchison, J., McCormack, C.G. & Sutherland, W.J. (2017) Some Aspects of Enhancing Natural Pest Control. Pages 359-381 in: W.J. Sutherland, L.V. Dicks, N. Ockendon & R.K. Smith (eds) What Works in Conservation 2017. Open Book Publishers, Cambridge, UK