Action

Use pesticides only when pests or crop damage reach threshold levels

How is the evidence assessed?
  • Effectiveness
    39%
  • Certainty
    30%
  • Harms
    20%

Study locations

Key messages

Natural enemies: One randomised, replicated, controlled study from Finland found that threshold-based spraying regimes increased numbers of natural enemies in two of three years but effects lasted for as little as three weeks.
Pests and disease: Two of four studies from France, Malaysia and the USA reported that pests were satisfactorily controlled. One randomised, replicated, controlled study found pest numbers were similar under threshold-based and conventional spraying regimes and one study reported that pest control was inadequate. A randomised, replicated, controlled study found mixed effects on disease severity.
Crop damage: Four of five randomised, replicated, controlled studies from New Zealand, the Philippines and the USA found similar crop damage under threshold-based and conventional, preventative spraying regimes, but one study found damage increased. Another study found slightly less crop damage compared to unsprayed controls.
Yield: Two of four randomised, replicated, controlled studies found similar yields under threshold-based and conventional spraying regimes. Two studies found mixed effects depending on site, year, pest stage/type or control treatment.
Profit: Two of three randomised, replicated, controlled studies found similar profits using threshold-based and conventional spraying regimes. One study found effects varied between sites and years.
Costs: Nine studies found fewer pesticide applications were needed and three studies found or predicted lower production costs.

Crops studied were barley, broccoli, cabbages, cauliflower, celery, cocoa, cotton, grape, peanut, potato, rice, tomato and wheat.

About key messages

Key messages provide a descriptive index to studies we have found that test this intervention.

Studies are not directly comparable or of equal value. When making decisions based on this evidence, you should consider factors such as study size, study design, reported metrics and relevance of the study to your situation, rather than simply counting the number of studies that support a particular interpretation.

Supporting evidence from individual studies

  1. A trial in Yavan, Tajikistan (Umarov & Tajibayev 1983) reported that spraying regimes based on thresholds of economic damage by pests favoured the build-up of natural enemy numbers in cotton Gossypium sp. crops. Up to 1,000 natural enemy insects were recorded per 100 plants in some instances, but more typical numbers were not presented. Farms using the threshold-based spraying regime applied pesticides to only 25-30% of cropped land, reportedly much less than on neighbouring farms. This conference abstract provided no details of experimental study design or the pesticide used.

    Study and other actions tested
  2. A study in 1982-1983 in arable land on Long Island, New York, USA (Leach et al. 1986) found potato Solanum tuberosum damage was similar for growers who sprayed insecticide according to threshold-based recommendations and growers who did not. Growers using a threshold-based spraying regime used an average of 1.4-2.6 fewer sprays/year for controlling Colorado potato beetle Leptinotarsa decemlineata and 0.4 fewer sprays/year for controlling aphids (Aphidoidea), compared to growers not using the regime. The threshold-based regime saved US$70-130/ha on Colorado potato beetle control and US$12/ha on aphid control. The trial included 30 growers, who were classified as having followed the recommendations if their practice matched recommendations 60% of the time. This included spraying within 72 hours of a recommendation to spray, and not spraying when not recommended. Growers used a variety of recommended insecticides including aldicarb, disulfoton, phosmet, fenvalerate, parathion and methamidophos. Effects on natural enemies were not presented.

    Study and other actions tested
  3. A randomised, replicated, controlled study in 1984-1985 in Florida, USA (Stanley et al. 1988) found more armyworm Spodoptera eridania damage to tomato Solanum lycopersicum (affecting 1.7-3.4% of fruits) in plots sprayed when pests exceeded threshold levels compared to plots sprayed weekly (0.7-0.9%). Yield was similar between plots receiving the threshold-based spraying regime (464-541 marketable fruits/10 plants) and plots receiving weekly sprays (470-600 marketable fruits). The threshold-based spraying regime used seven applications of insecticide on average compared to 14 applications in the weekly regime. In the former, sprays were applied when monitoring found at least 0.7 leafminer Liriomyza trifolii larvae on tomato leaflets, or at least one armyworm on fruiting tomato plants (prior to fruiting the threshold was one armyworm/six plants). Cyromazine and methamidophos insecticides were used to manage leafminers and fenvalerate and permethrin were used to manage armyworms. Each spraying regime was replicated 12 times and tested under different planting densities (4.5, 9.0 and 18.0 feet between rows). Effects on natural enemies were not presented.

    Study and other actions tested
  4. A randomised, replicated, controlled study in 1988-1989 in Oklahoma, USA (Roberts & Cartwright 1991) found similar numbers of cabbage loopers Trichoplusia ni, thrips Frankliniella spp. and aphids (Aphidoidea) in plots with a threshold-based spraying regime and plots with a conventional, weekly spraying regime, when averaged across the season. Damage to cabbages Brassica oleracea and cabbage yield were also similar in plots with threshold-based and conventional spraying regimes. Cabbages in the threshold-based regime were sprayed with Dipel 2X biological insecticide when moths and butterflies (Lepidoptera) averaged 0.5 larvae/plant in early to mid-growth stages and 0.3 larvae/plant in late growth stages. In the conventional regime plots received weekly sprays of Bacillus thuringiensis biological insecticide. Plots sprayed when pest thresholds were exceeded received 7-8 sprays compared to 11 sprays in the plots treated weekly. Each spraying regime was replicated 18 times. Insect pests were sampled on 5-10 plants/plot and one or two times/week. Figures and effects on natural enemies were not presented.

    Study and other actions tested
  5. A trial in 1985-1988 in Sabah, Malaysia (Wood et al. 1992) found that lindane applications, made when monitoring confirmed that cocoa pod borers Conopomorpha cramerella were present, did not stop infestations from continuing to increase in January-June 1986. In 1987-1988, cocoa Theobroma cacao pod infestation was similar when insecticide applications were determined using either low or high pod infestation thresholds. At peak levels in 1987 and 1988, 74% and 21% of pods were infested in the low-threshold plot, respectively, versus 84% and 17% in the high-threshold plot. In January-June 1986 cocoa pod borer moths were monitored using pheromone traps and lindane was applied seven times in response to positive catches. In 1987-1988, thresholds were set using an index of pod infestation that quantified the percentage of infested pods at four severity levels. An experimental field was divided into two plots which were assigned different thresholds (index values of 5 and 30) for applying insecticide. Low-threshold plots received more cypermethrin applications than high-thresholds plots (22 vs. seven applications in 1987, two vs. zero applications in 1988). Pod infestation was assessed for each threshold treatment using 200-400 pods taken from 3-4 randomly selected heaps of harvested crop. Effects on natural enemies were not presented.

    Study and other actions tested
  6. A randomised, replicated, controlled study in cereal fields in 1992-1994 in Finland (Huusela-Veistola 1998) found more spiders (Araneae) in plots using a threshold-based spraying regime (peaking at 17-31 spiders/3 traps) compared to conventional plots sprayed annually (12-23 spiders) in 1992 and 1994. The effect lasted for three weeks in 1992 and at least six weeks in 1994, but overall spider numbers were similar between treatments in 1993. More money spiders (Linyphiidae) were found in plots in the threshold-based spraying regime (peaking at 10-20 spiders/3 traps) than conventional plots (9-12 spiders) in all years. Wolf spider (Lycosidae) numbers were only greater in the threshold-based than conventional plots in 1994. At the species level, only one of three species tested (the money spider Erigone atra) was affected by pesticide regime type. In the threshold-based regime, sprays were made when control thresholds were exceeded, resulting in one insecticide (pirimicarb) spray in 1992 and one herbicide spray in 1994. Insecticides (dimethoate and deltamethrin), fungicides (carboxin, imazalil and propiconazole), herbicides and growth regulators were applied annually in conventional plots. Barley Hordeum vulgare was grown in 1992-1993 and wheat Triticum aestivum in 1994. Spiders were captured using pitfall traps monitored weekly (8-10 times between sowing and harvest).

    Study and other actions tested
  7. A randomised, replicated, controlled study in 1994-1997 in California, USA (Reitz et al. 1999) found pest damage on celery Apium graveolens was similar in plots receiving threshold-based insecticide applications (5-39% plants damaged) and conventionally treated plots (5-33%) in 1995-1997. In 1994, damage was greater in threshold-based (38% plants) than conventional (20%) plots. Net profit was similar between threshold-based and conventional plots in 1994-1995 and 1997. In 1996 (an unprofitable year) net loss was smaller in threshold-based than conventional plots. A separate randomised, replicated, controlled commercial trial in 1997 found similar yield and net profit from plots with threshold-based applications (1,105-1,121 marketable cartons and US$8,000-8,330) and conventional plots (1,104 cartons and US$8,330). In the 1994-1997 test, threshold plots received selective insecticides (3-4 applications/year) when pest insect thresholds were exceeded. Conventional plots received broad spectrum insecticides to prevent pest build-up (8-9 applications/year). Treatments were replicated four times in plots of 16 celery rows, 20 m long. In 1997, conventional plots (receiving insecticide and fungicide) were compared with plots receiving either threshold-based insecticide application, threshold-based fungicide application, or both. Up to four fungicide and seven insecticide types were used per treatment. Treatments were replicated four times (0.4 ha plots). Effects on natural enemies were not presented.

    Study and other actions tested
  8. A replicated study in 1998-2001 in Gisborne and Hawke’s Bay, New Zealand (Walker et al. 2004) reported better control of pests by natural enemies when a threshold-based spraying regime was applied compared to conventional regimes. Insecticide use in vegetable Brassica spp. crops was reduced by 40-70% in the second year of the threshold-based spraying programme compared to conventional regimes. The threshold-based regime could potentially save NZ$125/ha when accounting for the cost of monitoring pests and assuming that spraying could be reduced by 3-4 applications/ha. Sprays (using a variety of insecticide types) were timed according to thresholds of diamondback moth Plutella xylostella and aphid (Aphidoidea) infestation, with selective insecticides used in rotation within each year. Details of experimental setup were not provided.

    Study and other actions tested
  9. A 13-year randomised, replicated, controlled study in the Philippines (Litsinger et al. 2005, the same study as Litsinger et al. 2006a and 2006b) found that spraying based on thresholds of pest abundance or damage typically resulted in less control of rice Oryza sativa leaf damage (averaging 31-34% control) than a full protection regime (60%), and similar control to preventative (41%) and farmers’ practice regimes (24%). Rice yields were lower in threshold-based spraying regimes (4.5-4.6 t/ha) than for full protection (5.0 t/ha), similar to preventative (4.8 t/ha) and farmers’ practice (4.5 t/ha) regimes but greater than in untreated plots (4.4 t/ha). Average monetary return from threshold-based spraying ranged from a US$-23/ha loss to a US$48/ha gain, compared to US$-34/ha to US$24/ha with the preventative regime and US$-9/ha to US$28/ha with farmers' practice. Thresholds were studied at 4-9 farms/year in four sites totalling 68 rice crops overall. Plots receiving full protection were sprayed with insecticides weekly. Plots receiving the preventative regime received carbofuran granules and two insecticide (chlorpyrifos) sprays. Farmers' practice involved low insecticide doses and timing based on prevention or very low pest thresholds. Plots measured 100-200 m². Leaf damage was measured relative to 0% control in untreated plots. Effects on natural enemies were not presented.

    Study and other actions tested
  10. A 13-year randomised, replicated, controlled study in the Philippines (Litsinger et al. 2006a, the same study as Litsinger et al. 2005 and 2006b) found that spraying based on thresholds of leaffolder (Crambidae and Pyralidae) abundance or damage typically resulted in greater rice Oryza sativa yields than unsprayed controls (averaging 41-263 kg/ha greater). Yields were improved with five out of eight different spraying thresholds tested during initial rice growth, and with six out of eight thresholds tested in the flowering and ripening stages. The most effective threshold was that of 10-15% leaffolder leaf damage, which enabled correct decisions to spray/not spray on 93-99% of occasions. Thresholds based on the numbers of leaffolder larvae or moths led to 91-100% and 62-96% correct decisions, respectively. Synthetic insecticides (endosulfan, monocrotophos, BPMC and azinphos-ethyl) controlled 60-100% of leaffolder larvae but only 13-53% of leaf damage. Thresholds were studied at 4-9 farms/year in each of four sites across the Philippines, totalling 68 rice crops overall. Thresholds and unsprayed controls were tested in 100-200 m² plots. Decisions were considered correct if leaf damage reached 10-15% and yield loss exceeded 250 kg/ha (in additional control plots). Comparisons to a conventional spraying regime and effects on natural enemies were not presented.

    Study and other actions tested
  11. A 13-year randomised, replicated, controlled study in the Philippines (Litsinger et al. 2006b, the same study as Litsinger et al. 2005 and 2006a) found greater rice Oryza sativa yield (by 130-299 kg/ha) when sprays were based on thresholds of stem borer Scirpophaga spp. egg abundance and rice plant damage compared with no treatment. Thresholds based on moth abundance resulted in similar yields to untreated controls (17-65 kg/ha difference). The most effective thresholds assessed deadhearts (percentage of leaves killed by stem borers) and enabled correct decisions to spray/not spray on 90-99% of occasions. The most effective threshold level depended on the stage of crop growth (5%, 25% and 10% deadhearts for initial, flowering and ripening stages, respectively). At best, insecticides (endosulfan, chlorpyrifos or chlorpyrifos and BMPC) controlled 36% of plant damage and at worst they increased damage by 5% (at three weeks after planting). Thresholds were studied at 4-9 farms/year in each of four sites, totalling 68 rice crops overall. Thresholds and unsprayed controls were tested in 100-200 m² plots. Decisions were considered correct if deadhearts reached 5-15% per 20 plant hills, and if yield loss exceeded 250 kg/ha (in additional control plots). Comparisons to a conventional spraying regime and effects on natural enemies were not presented.

    Study and other actions tested
  12. A randomised, replicated, controlled study in 2002-2004 in Georgia, USA (Nuti et al. 2008) found more severe stem rot Sclerotium rolfsii in sprayed plots treated when weather and soil temperature thresholds were met than in conventional plots treated regularly (10-22 vs. 5-8 average disease rating), on two occasions. Stem rot severity was similar between plots on three other occasions. Leaf spot (Mycosphaerellaceae) severity showed inconsistent differences between threshold-based and conventional plots (1.3-4.3 vs. 1.9-3.9 average disease ratings). Yields were similar in threshold-based and conventional plots on three occasions (3,290-4,750 vs. 3,260-4,980 kg/ha), but lower in threshold-based than conventional plots on two occasions (4,200 vs. 4,560 kg/ha and 3,040 vs. 3,490 kg/ha). Monetary return was similar between threshold-based and conventional plots on three occasions (US$1,050-1,750/ha vs. US$1,110-1,770/ha), but lower in threshold-based plots on one occasion (US$1,480 vs. US$1,580). Fungicide regimes were tested in peanut Arachis hypogea crops for three years at two sites (up to six occasions in total). Rainfall thresholds were used to time fungicide sprays, and soil temperature thresholds were used to select between tebuconazole and chlorothalonil fungicides. Weather and soil temperature closely related to types and extents of fungal diseases. Effects on natural enemies were not presented.

    Study and other actions tested
  13. A trial in 2000-2002 in Hawke’s Bay, New Zealand (Walker et al. 2010) found that tomato Solanum lycopersicum damage did not exceed the commercially acceptable level of 5% at harvest on 16 of 17 occasions when decisions to spray were made using thresholds of cotton bollworm Helicoverpa armigera larvae abundance. On average, 0.8-5.5% of tomatoes were damaged in treated fields compared with 3.9-7.1% in unsprayed fields (where thresholds were also exceeded). The decision to not treat crops while maintaining < 5% damage was correct 10 out of 11 times. Treatment decisions were made for 22 fields which all met or exceeded a conventional threshold of one larvae/plant. However, in 16 fields and one half-field, spraying only took place if larvae numbers exceeded an adjustable threshold accounting for site-specific parasitism rates (see 'Incorporate parasitism rates when setting thresholds for insecticide use'). Thresholds varied from 1-8.3 larvae/plant. Insecticides included spinosad or Bacillus thuringiensis pesticidal bacteria. Fruit damage was assessed for 40 random plants/field. Three fields and one half-field were left unsprayed to assess crop damage without treatment, but the effects of spraying with and without a pest threshold approach were not compared. Effects on natural enemies were not presented.

    Study and other actions tested
  14. A trial in 2008-2009 in Aquitaine, France (van Helden et al. 2011) during flavescence dorée outbreaks reported that a threshold-based spraying regime controlled this disease (caused by Candidatus Phytoplasma vitis bacteria) in two grape Vitis vinifera vineyard landscapes, as the disease carrying pest (the American grapevine leafhopper Scaphoideus titanus) was very scarce after the first insecticidal spray. In one landscape, the threshold-based regime resulted in 4,299 sprayings compared to an estimated 14,919 sprayings had a conventional approach been taken. Spraying frequency was reduced by 54-72% across the two landscapes, and spraying costs fell by €19-29/ha. Compared to the conventional approach of three insecticide applications, the threshold-based regime comprised one application, followed by a second spray if leafhopper abundance exceeded 3 adults/trap. One yellow delta trap was set up per 30 ha (costing approximately €3/ha) and checked weekly. The threshold-based regime used circular buffers (with a 2 km radius) to define spraying areas around infected sites whereas, traditionally, entire districts were sprayed under the conventional regime. The threshold-based spraying regime was applied in two landscapes, and compared with theoretical estimates for insecticide use under the conventional regime. Effects on natural enemies were not presented and the insecticide type was not specified.

    Study and other actions tested
Please cite as:

Wright, H.L., Ashpole, J.E., Dicks, L.V., Hutchison, J., McCormack, C.G. & Sutherland, W.J. (2020) Some Aspects of Enhancing Natural Pest Control. Pages 589-612 in: W.J. Sutherland, L.V. Dicks, S.O. Petrovan & R.K. Smith (eds) What Works in Conservation 2020. Open Book Publishers, Cambridge, UK

Where has this evidence come from?

List of journals searched by synopsis

All the journals searched for all synopses

Natural Pest Control

This Action forms part of the Action Synopsis:

Natural Pest Control
Natural Pest Control

Natural Pest Control - Published 2013

Natural Pest Control Synopsis

What Works 2021 cover

What Works in Conservation

What Works in Conservation provides expert assessments of the effectiveness of actions, based on summarised evidence, in synopses. Subjects covered so far include amphibians, birds, mammals, forests, peatland and control of freshwater invasive species. More are in progress.

More about What Works in Conservation

Download free PDF or purchase
The Conservation Evidence Journal

The Conservation Evidence Journal

An online, free to publish in, open-access journal publishing results from research and projects that test the effectiveness of conservation actions.

Read the latest volume: Volume 21

Go to the CE Journal

Discover more on our blog

Our blog contains the latest news and updates from the Conservation Evidence team, the Conservation Evidence Journal, and our global partners in evidence-based conservation.


Who uses Conservation Evidence?

Meet some of the evidence champions

Endangered Landscape ProgrammeRed List Champion - Arc Kent Wildlife Trust The Rufford Foundation Save the Frogs - Ghana Mauritian Wildlife Supporting Conservation Leaders
Sustainability Dashboard National Biodiversity Network Frog Life The international journey of Conservation - Oryx Cool Farm Alliance UNEP AWFA Bat Conservation InternationalPeople trust for endangered species Vincet Wildlife Trust