Crop damage by vertebrates in Latin America: current knowledge and potential future management directions

Compilation of studies

The searches returned 118 records that met all initial selection criteria, and an additional four records previously known to the authors were later included. Of the 122 records, seven were excluded because they were gray literature: six were MSc or Ph.D. theses, and one was a technical report. Two additional records were excluded because they studied damage by vertebrates to silo bags (Zufiaurre, Abba & Bilenca, 2020) and farming machinery (Álamo Iriarte, Sartor & Bernardos, 2019) and not to crops directly. After this process, 113 studies were included in the final analysis (Fig. 1).

Geographic and temporal distribution of studies

The temporal distribution of studies showed that there has been an increase in the number of studies published on the topic since the 1980s. Only 21 studies (19%) were published before 2000 and 64 (57%) were published in the last decade (2011–2020). The year with the most studies published was 2018, with 11 (10%) studies (Fig. 2). The studies were scattered across most Central and South America, with some in the Caribbean (Fig. 3). The study sites were located across areas with different proportions of land cover by crops (Fig. 3). The countries with the highest number of studies were Brazil (n = 31), Argentina (n = 18), and Mexico (n = 13). Other countries with study sites included Costa Rica (n = 10); Peru (n = 9); Venezuela (n = 8); Bolivia, and Uruguay (n = 7); Colombia (n = 6); Barbados (n = 3); Cuba, Chile, and Puerto Rico (n = 2); and Ecuador, Guyana, Belize, Dominican Republic, and Saint Kitts and Nevis (n = 1) (Table S1). Four studies were conducted in more than one country.

Type of crop plantations and studies

The majority of the studies were conducted on commercial plantations (n = 60, 53%), followed by subsistence or semi-subsistence plantations (n = 23, 20%). The remaining studies were conducted in experimental fields (n = 6), laboratories (n = 2), harvesting concessions (n = 1), and areas for which the type of plantation could not be reliably determined (n = 22).

Among the six categories in which the studies were categorized based on their focus, the one with the highest number of studies was “Crop damage evaluation” with 42 studies (37%), followed by “Farmer perception” with 38 studies (34%), and “Crop-feeding species behavior” with 37 studies (33%). The other three categories had fewer studies: “Pest species or outbreak overview” included 16 studies (14%), “Crop protection experiment” had 10 studies (9%), and “Protection technique evaluation” included six studies (5%) (Table S1).

Vertebrates and crops

In total, 272 crop-damaging vertebrate taxa were studied in 113 reviewed studies; these included mammals, birds, and reptiles (Table S2). The number of taxa included in each study varied greatly, ranging from 1 to 56, with 64 (57%) studies focusing on a single vertebrate species. The mammal taxa represented nine different orders: Rodentia (70 taxa), Primates (12 taxa), Carnivora (11 taxa), Cingulata (8 taxa), Artiodactyla (7 taxa), Didelphimorphia (6 taxa), Lagomorpha (4 taxa), Perissodactyla (2 taxa), and Chiroptera (1 taxon). The bird taxa represented 13 orders: Passeriformes (77 taxa), Psittaciformes (22 taxa), Columbiformes (18 taxa), Anseriformes (11 taxa), Piciformes (7 taxa), Gruiformes (4 taxa), Galliformes (3 taxa), Accipitriformes (2 taxa), Charadriiformes (2 taxa), Cariamiformes (1 taxon), Cuculiformes (1 taxon), Pelecaniformes (1 taxon), and Strigiformes (1 taxon). Finally, there was only one reptile taxon in the order Squamata (Table S2). The most represented order among the reviewed studies was Rodentia, with an importance value of 126, followed by Passeriformes (123), Columbiformes (55), Psittaciformes (37), Carnivora (30), Artiodactyla (26), Primates (20), and Anseriformes (15) (Fig. 4). Together, these eight orders accounted for almost 90% of the representation in all studies. The remaining orders had importance values below 10 (Fig. 4).

Across all reviewed studies, 77 genera of crops were reported to be damaged by vertebrates, with the number of genera per study varying from 1 to 31. Forty-six studies included only one crop genus, and 12 did not specify which crops were affected by vertebrates. The most prominent crop in the studies was corn (Zea sp.). It was featured in 55 (49%) studies and interacted with 17 of the 23 vertebrate orders represented in our review, most predominantly with Rodentia, Psittaciformes, and Artiodactyla (in 18, 16, and 15 studies, respectively) (Fig. 5). The second most represented crop was rice (Oryza sp.), which appeared in 28 (23%) studies and was mainly damaged by Passeriformes and Rodentia (11 studies each) (Fig. 5). Sorghum (Sorgum sp.) reportedly suffered damage in 23 (20%) studies, mainly by three bird orders: Columbiformes (nine studies), Passeriformes and Psittaciformes (six studies each). Sugarcane (Saccharum sp.) was mentioned in 19 (17%) studies and interacted with four mammalian orders: Rodentia (9), Primates (seven studies), and Artiodactyla and Carnivora (two studies each) (Fig. 5). Both, soy (Glycine sp.) and banana (Musa sp.), were mentioned in 15 (13%) studies. Soy was damaged mostly by Columbiformes (seven studies), and Artiodactyla and Psittaciformes (three studies each) (Fig. 5). Bananas interacted the most with Primates (six studies), Carnivora (5), and Rodentia (4) (Fig. 5). Wheat (Triticum sp.), beans (Phaseolus sp.), and sunflowers (Helianthus sp.) were mentioned in 14 (12%) studies. Damage to wheat has been reported mainly by Columbiformes (seven studies). Beans suffered damage mainly by Rodentia (six studies). Sunflowers mostly interacted with Psittaciformes (eight studies) and Columbiformes (5). Finally, manioc (Manihot sp.) was included in 12 (11%) studies and was damaged mainly by Artiodactyla and Rodentia (six studies each) (Fig. 5).

Of the 238 vertebrate taxa that could be identified at the species level from our review, 22 were not categorized as Least Concern (LC), four species considered Endangered (EN), eight species considered Vulnerable (VU), seven species categorized as Near Threatened (NT), and three as Data Deficient (DD) (IUCN, 2021). Of these species, four were birds, three were psittacines (VU – Eupsittula canicularis, NT – Amazona aestiva and Eupsittula nana), and one was from the Galliformes order (NT – Colinus virginianus). The remaining species not categorized as LC were mammals: eight primates (EN – Leontopithecus chrysomelas and Sapajus flavius, VU – Alouatta guariba, Alouatta palliata and Cebus capucinus, NT – Sapajus libidinosus, Sapajus nigritus, and Erythrocebus patas), four rodents (EN – Callistomys pictus, VU – Oryzomys laticeps, DD – Dasyprocta variegata and Galea musteloides), one carnivore (VU – Tremarctos ornatus), two even-toed ungulates (VU – Tayassu pecari and DD – Mazama americana), two odd-toed ungulates (EN – Tapirus bairdii and VU – Tapirus terrestris), and one cingulate (NT – Dasypus hybridus) (IUCN, 2021).

Protection techniques

In the reviewed studies, more than half (n = 66, 58%) tested or mentioned a range of diverse techniques used to protect crops from vertebrates (Table S3). The most frequently used control method was hunting (either with weapons, dogs, or traps), which was mentioned in 37 (56%) of the studies that mentioned protection techniques (Fig. 6). The second most widely represented technique was another type of lethal control, the use of poisons, which was reported in 22 (33%) studies (Fig. 6). The following poisonous substances have been reported: herbicides, rodenticides, organophosphates, sodium monofuroacetate, coumarin, pyriminil, diphacinone, biorat, estricinina, methyl bromide, metomil, aluminum phosphate, zinc phosphide, parathion, chlorpyrifos, monocrotophos, endrin, mevinphos, dicrotophos, CPT, CPTH, thallium sulfate, coumatetralyl, brodifacoum, thiodicarb, and carbofuran. Agricultural practices were mentioned in 20 (30%) studies, including field clearing, time of planting or harvest, changing the location or the type of crops, altering the density of the crops, using barrier crops or firebreaks, and providing alternative food sources. Acoustic deterrents were reported in 15 (23%) studies (Fig. 6), including firecrackers, gas cannons, firearms, yellings, sirens, predator sounds, distress calls, and horns. Visual deterrents were used in 14 (21%) studies (Fig. 6) and consisted of scarecrows, reflective objects, smoke, fire, flags, predator outlines, balloons, calcium carbonate paint, and carpenter’s chalk. Chemical repellents, including anthraquinone, methiocarb, methyl anthranilate, bidrim, thrimethacarb, dimethyl, methyl anthranilate, synergized aluminum, ammonium sulfate, copper oxalate, copper oxychloride, condensed tannins, avitrol, and soap, were reported in 12 (18%) studies (Fig. 6). Vigilance by people or guard dogs was mentioned in 12 (18%) studies. Physical barriers, such as nets, fences, electric fences, trenches, metal bands, and wire mesh, were included in eight (12%) studies (Fig. 6). Biological control in the form of introducing infectious diseases, introducing or attracting predators, or reducing suitable habitats was mentioned in six (9%) studies (Fig. 6). Reproductive control and olfactory repellants were used in four (6%) studies (Fig. 6). The types of reproductive controls mentioned were the use of sterilants, nest burning, and egg destruction. The olfactory repellents mentioned were creolin, Tabebuia extract, burnt rubber, and human odor. Capture and relocation were mentioned in two (3%) studies (Fig. 6). Finally, Capsicum was used as a palatable deterrent in one (2%) study (Fig. 6).

Quality of the information reported

The methodology used to identify vertebrate taxa responsible for crop-feeding and to quantify the damage to crops varied greatly among studies. The most common identification methods used were direct observation and interviews with farmers (n = 40 and 36 studies, respectively), followed by the interpretation of indirect signs (n = 24). The least used methods were trapping or hunting (n = 14), looking at stomach or crop contents (n = 8), using of previous knowledge (n = 9), expert reports (n = 7), camera traps and radiotelemetry (n = 3 each), using distribution maps and museum specimens (n = 2 each), and using stable isotope analyses, farmer complaints and hunter reports (n = 1 each). The most common method used to quantify crop damage was measuring the proportion of damaged crops (i.e., area, plants, fruits, and production), which was used in 37 studies. Interviews with local farmers were used to estimate the crop damage in 22 studies. Eight studies estimated the economic cost of crop destruction, six examined stomach or crop contents, and two studies used models to predict damage.

Of the 66 studies that mentioned the use of protection techniques, 25 indicated their effectiveness, while the other 41 studies only listed or alluded to the damage control methods used in their settings (Fig. 6, Table S3). Of the 25 studies that evaluated the effectiveness of the protection techniques, only 10 performed experiments and reported their results. The remaining studies either conveyed effectiveness by asking farmers in surveys (eight studies), or through author discussions of the effectiveness of the control methods (seven studies).

Geographic and temporal distribution of studies

Our results showed that crop damage by vertebrates in Central and South America did not receive much attention in the published literature before 2000, with most studies being published after 2011. Considering the overall 2.3 fold increase in scientific literature worldwide from 1,067,910 articles in 2000 to 2,554,373 articles in 2018 (World Bank Data, 2021b), the number of articles on vertebrate crop damage in Latin America is increasing rapidly. Despite this increase in published articles, it is still an emerging discipline considering the projections of cropland expansion (Riahi et al., 2017), and the fact that South America has been the world’s leading region in cropland cover expansion over the past two decades (Popkin, 2022).

The country with the highest number of studies was Brazil, which is likely a consequence of its large territorial area and extensive research. Brazil is the largest country in Central and South America and the largest producer of studies in scientific and technical journals (60,148 in 2018, placing 18th worldwide) (World Bank Data, 2021b). Another important factor may be its crop production rates; data for crop production (cereals fruits, vegetables, sugar crops, roots and tubers, tree nuts, fiber crops, and oil crops) in 2019 reveals that Brazil has the highest crop production rates among all Latin American countries (FAO, 2021). Argentina and Mexico were the next two countries with the highest number of studies; both are also large countries with high scientific and crop production rates (FAO, 2021; World Bank Data, 2021b). Studies in these two countries have focused on specific vertebrate groups. All but two of the studies conducted in Argentina focused on crop damage by birds, with two species being the most frequent: the eared dove (Zenaida auriculata) and the monk parakeet (Myiopsitta monachus). Nearly half of the studies from Mexico have focused on agricultural rodent pests.

The Costa Rican case is interesting because it has a small territorial area and scientific production compared with other countries in Central and South America (only 507 articles were published in 2018) (World Bank Data, 2021b). It was also placed near the bottom of almost all crop categories evaluated in 2019 (FAO, 2021). Despite this, a relatively large number of studies have been published on crop damage by vertebrates from the country. Thus, our review indicated a higher interest in research on crop-feeding by vertebrates in Costa Rica than in other Latin American countries.

It is important to highlight that the number of published studies in a country is not directly proportional to the corresponding severity of the crop damage by vertebrates, as many smaller countries have little-to-none scientific production but have the conditions to potentially be severely affected by this type of human-wildlife conflict. Countries such as the Guianas, most Central American countries, and many Caribbean island nations have few or no published studies on the topic, but are rich in wildlife biodiversity (Mittermeier et al., 2011; Myers et al., 2000), and have high rates of poverty (Fisher & Christopher, 2007; World Bank Data, 2021a). Crop-feeding is an ecosystem disservice derived from high biodiversity (Ango et al., 2014; Naughton-Treves et al., 2003), and it can damage the economy of vulnerable communities (Gontse, Mbaiwa & Thakadu, 2018; Raphela & Pillay, 2021). The combination of these factors increases the risk of conflict between biodiversity conservation and food security (Molotoks et al., 2017). Therefore, future research on crop-feeding should focus on these countries.

Type of plantations and studies

More than half of the studies in our review were concentrated in commercial crop plantations. This could be expected because commercial plantations generally have much larger areas (Felix et al., 2014; Lima et al., 2019; Lobão & Nogueira-Filho, 2011) than smaller subsistence plantations (Can-Hernández et al., 2019; Chaves & Bicca-Marques, 2017; Naughton-Treves et al., 2003). Moreover, crop losses on subsistence plantations tend to be better tolerated by landowners, as their main objective is not linked to profit (Chaves & Bicca-Marques, 2017; Rocha & Fortes, 2015; Spagnoletti et al., 2017). Subsistence farmers also tend to be more knowledgeable about the wildlife that inhabits their fields, and they engage in practices that are beneficial to conservation (Silva-Andrade et al., 2016). However, there are also examples of communities that engage in hunting to defend their subsistence crops (Can-Hernández et al., 2019; Cossios, Ridoutt & Donoso, 2018).

Of all reviewed studies, 37% evaluated the magnitude of crop damage by vertebrates, but only 14% focused on crop protection techniques. Moreover, a large proportion of studies used interviews with local farmers to collect data and evaluate their perceptions. This method was the second most commonly used method for identifying crop-feeding vertebrates and quantifying crop damage. However, interviews were only corroborated by alternative methods in a few studies, which could present an inherent bias in the reported crop damage. Involving local communities and stakeholders in research can have positive effects on nature conservation (Beierle & Konisky, 2001; Young et al., 2013) and enable data collection over large areas (Michalski et al., 2020; Michalski & Peres, 2017). Farmers’ perceptions and knowledge are central to studies on crop damage and conservation strategies. However, relying solely on farmers’ perception of crop damage can be misleading, as their ideas of which species are responsible for damaging crops or the extent of losses may not accurately represent the reality (Albarracín & Aliaga-Rossel, 2018; Flores-Armillas et al., 2020; Hill, 2004) or be proportional to the scale of the problem (Simonsen, Tombre & Madsen, 2017). Therefore, relying almost exclusively on interviews with local farmers for data generation may result in an incorrect assessment of the conflict, which when coupled with an exaggerated perception of damages caused by vertebrates may lead to an increase in the use of lethal methods for retaliation (Can-Hernández et al., 2019). Studies that perform field validation of crop damage are important, and more effort towards using field validation methods must be made in future studies.

Vertebrates and crops

Among the 272 vertebrate taxa that were identified in the studies as causing crop damage, Rodentia was the order of vertebrates with the highest importance value. Rodents have long been considered as some of the worst crop pests worldwide (Capizzi, Bertolino & Mortelliti, 2014; Lauret et al., 2020; Stenseth et al., 2003). This concurs with our results, where they were shown to cause damage to the highest number of crop genera (45), affecting corn the most (Felix et al., 2014; Ferraz et al., 2003). Early studies on crop damage and pest control in Latin America focused on rodents (Espinoza & Rowe, 1979), and they have continued to be the main focus of research during the period included in our review (Felix et al., 2014; Ferraz et al., 2003; Sánchez-Cordero & Martínez-Meyer, 2000; Santos, 2018). Rodents can cause extensive damage to crops and have often been the target of lethal control (Hilje, 1992; Villafaña Martín et al., 1999). Four rodent species (paca – Cuniculus paca, capybara – Hydrochoerus hydrochaeris, and hispid cotton rat – Sigmodon hispidus, and the black rat – Rattus rattus) appeared in the largest number of studies.

The second order of mammals with the highest importance values was Carnivora, which interacted with 21 crop genera. Three species (Nasua narica, Nasua nasua and Procyon lotor), all belonging to the Procyonidae family, have been recorded in several studies. These species are often among the most concerning to farmers (Castillo-Chinchilla et al., 2018) and among the most damaging to crops, particularly to corn (Can-Hernández et al., 2019; Flores-Armillas et al., 2020).

The Artiodactyla order affects 18 different crop genera, but interacts mostly with corn and manioc, often causing extensive damage (Abrahams, Peres & Costa, 2018; Pérez & Pacheco, 2014; Romero-Balderas et al., 2006). Among the even-toed ungulates, two species (collared peccary – Pecari tajacu and wild boars – Sus scrofa) had the highest number of appearances in all studies. Wild boars are invasive in many parts of the world, including in Latin America, and cause extensive crop damage worldwide (Bevins et al., 2014). They can have deleterious effects on native biodiversity around the globe, even driving some species to extinction (Risch, Ringma & Price, 2021). In our review, all studies focusing on wild boars were from Brazil, where boars have been found to dominate local communities shortly after the invasion (Doutel-Ribas et al., 2019) and consume large amounts of cultivated grain (Cervo & Guadagnin, 2020). Lethal methods for wild boar control have been legal in the country since 2013, and hunting has become widespread (Rosa, Wallau & Pedrosa, 2018). Most farmers agree that this species should be eradicated (Pereira, Rosa & Zanzini, 2019).

Primates were the order that interacted with the second-highest number of crop genera (43), with corn being the top crop interaction followed by sugarcane and bananas. Primates feeding on crops were often perceived as tolerable by farmers, and they rarely used lethal control measures against them (Chaves & Bicca-Marques, 2017; Lins & Ferreira, 2019; McKinney, 2019; Rocha & Fortes, 2015; Spagnoletti et al., 2017). This might be due to them often targeting crops that are not used commercially, which could favor a peaceful coexistence between humans and non-human primate crop-feeders (Chaves & Bicca-Marques, 2017; Rocha & Fortes, 2015; Spagnoletti et al., 2017). The tolerance of crop-feeding by primates might have also been motivated by their resemblance to humans, which causes empathy (Dore, Eller & Eller, 2018; Rocha & Fortes, 2015). Lethal control of primates was only recorded in the case of invasive vervet monkeys (Chlorocebus aethiops) in Barbados, where they cause damage to a variety of crops, and campaigns to reduce their population have been conducted (Boulton, Horrocks & Baulu, 1996). Lethal control of primates is also infrequent in Africa and Asia, where most farmers use non-lethal techniques (Marchal & Hill, 2009; Mc Guinness & Taylor, 2014; Siljander et al., 2020).

Like rodents, birds have long been considered agricultural pests, and the damage they cause to crops is a global concern (Anderson et al., 2013; de Mey, Demont & Diagne, 2012; Kale et al., 2014; Montràs-Janer et al., 2019). These perceptions have often motivated lethal control methods in an effort to reduce bird populations; however, these attempts are often unsuccessful (Linz et al., 2015). Among the studies included in this review, the trend of negative perceptions by farmers and their usage of lethal or reproductive control has continued (Basili & Temple, 1999; Bucher & Ranvaud, 2006; Canavelli, Swisher & Branch, 2013), although in some cases, non-lethal protection techniques have been tested with positive results (Avery, Tillman & Laukert, 2001; Robles et al., 2003). Some studies have found that bird species that feed on crops, such as sheldgeese (Chloephaga sp.) or mourning doves (Zenaida macroura), offset their negative impact by also feeding on weeds, which benefits crop production (García & Peiró, 2016; Gorosábel et al., 2019).

Among birds, Passeriformes had the highest importance value, being the second most recorded order after Rodentia. Despite this, the only passerine species that appeared in more than four studies was the house sparrow (Passer domesticus). Passerines interacted with 26 crop genera, but most of the damage was concentrated in rice, corn, and sorghum. Columbiformes have been reported to cause frequent damage to a wide range of crop genera, including corn, sorghum, wheat, soy, rice, and sunflowers. Three of the most prominent crop pests in Latin America belong to this order: the eared dove (Zenaida auriculata), which appeared in 15 studies, and two species of pigeons (Patagioenas maculosa and P. picazuro), which cumulatively appeared in 11 studies. Damage by Psittaciformes was concentrated mostly on corn, followed by sunflower and sorghum. Another one of the main bird pest species on the continent is the psittacine monk parakeet (Myiopsitta monachus), which was reported in 11 studies. These three pests (doves, pigeons, and parakeets) cause extensive damage to agricultural crops in many countries. However, studies on them have mostly been conducted in Argentina and Uruguay, where they have been the subject of many damage control methods (Bruggers, Rodriguez & Zaccagnini, 1998; Canavelli, Aramburú & Zaccagnini, 2012). Anseriformes were mostly reported to damage wheat and rice in their wintering areas, causing conflicts with local farmers (Gorosábel et al., 2019).

The order with the most threatened species was Primates. Thus, it is a good prospect for their conservation that farmers in Latin America tend to tolerate crop-feeding by primates and seldom use lethal control against them (Chaves & Bicca-Marques, 2017; Lins & Ferreira, 2019; McKinney, 2019). Most of the species that are not considered of least concern are infrequent in the literature, with the most frequent being the Andean bear (Tremarctos ornatus) and white-lipped peccary (Tayassu pecari), which appeared in four and three studies, respectively. Although not frequently cited in the revised literature, some of these species have been reported to cause extensive damage or are of great concern to farmers. For example, in a Peruvian study, the Brazilian tapir (Tapirus terrestris) was an infrequent crop-feeder but caused the largest proportion of damage per affected field, and it was hunted by locals to offset crop losses (Naughton-Treves et al., 2003). The cacao-rat (Oryzomys laticeps) caused the most damage and generated the highest number of complaints from farmers in the Brazilian Atlantic Forest, where farmers used lethal control methods against it (Lobão & Nogueira-Filho, 2011). White-lipped peccary (Tayassu pecari) causes damage to corn plantations in the Brazilian state of Mato Grosso, and farmers periodically cull the local population using firearms, traps, and mass poisoning (Lima et al., 2019). Lastly, the Andean bear (Tremarctos ornatus) caused little damage to banana and plantain crops in Colombia but generated strong negative attitudes among locals towards their presence and conservation efforts (Escobar-Lasso et al., 2020). The human-wildlife conflicts involving these threatened species may hinder conservation efforts by reducing the tolerance of local farmers and motivating lethal control.

Protection techniques

In our review, many types of crop protection techniques have been reported, but lethal control of crop-feeding populations via hunting and poisoning was the most commonly used protection method. Farmers may turn to lethal control after trying other protection techniques without success (Lima et al., 2019), and they tend to perceive hunting or poisoning as the most effective damage control method (Abrahams, Peres & Costa, 2018; Canavelli, Swisher & Branch, 2013; Lima et al., 2019). However, very few studies have provided reliable evidence that lethal control effectively reduces crop damage. Of the 37 studies that reported hunting as a control measure, only 10 evaluated its effectiveness, and only one managed to perform experiments. Pérez & Pacheco (2014) reported a reduction in crop damage (from 27.61% to 4.59%) in hunted crop fields when compared to control plots, but the effectiveness of hunting was only slightly higher than that of non-lethal alternatives (combination of agricultural practices, olfactory and visual deterrents, and vigilance).

The use of hunting as the main technique for reducing crop damage poses a series of problems. First, many of the most prevalent crop-feeders are species that have short life cycles and high reproductive rates, such as rodents, which enables them to recover faster from reductions in population size (Hein & Jacob, 2015). Hunting may also cause targeted species to modify their movement patterns and activity regimes (Béchet et al., 2003; Keuling, Stier & Roth, 2008; Little et al., 2016; McGrath, Terhune & Martin, 2018), which may alter the area and intensity of crop damage. Finally, trapping has been reported to be the most effective way of hunting to reduce vertebrate populations, but acquiring and maintaining traps can be expensive and labor-intensive (Rosa, Wallau & Pedrosa, 2018). Even in situations where hunting is not an effective method for protecting crops, it can provide farmers with alternative sources of food or income (Naughton-Treves et al., 2003) or grant social status (Cossios, Ridoutt & Donoso, 2018), which could explain its popularity as a protection technique among farmers.

The effectiveness of the use of poisons to control crop damage was evaluated in nine of the 22 studies that mentioned it. Six of the studies deemed that the use of poisons was effective in reducing crop damage, three of which performed experiments. Generally, poisons are considered an easy way to reduce populations of crop-feeding species, but their use can cause several environmental problems, such as non-target or vulnerable species being seriously harmed. Lima et al. (2019) reported that hundreds of white-lipped peccaries (Tayassu pecari) were killed simultaneously through the use of poisonous substances. Similarly, bird species that roost in large groups, such as Dickcissels (Spiza americana), can be killed in great numbers when their nesting sites or watering holes are poisoned (Basili & Temple, 1999). Furthermore, poisons can have severe consequences for carnivores and scavengers that feed on the carcasses of poisoned animals (Baudrot et al., 2020; Kalaivanan et al., 2011). In addition to the dangers that chemical pesticides pose to the environment, they can also pose a serious threat to the health of human workers and consumers (Rani et al., 2021).

To protect the environment and native wildlife alongside the interests of local communities, alternative methods for pest control must be tested and developed. From our review, only seven studies performed experiments to test the effectiveness of non-lethal crop protection techniques. Wire mesh exclosures significantly reduced damage by wildlife to manioc and walusa but not to corn in Bolivia (Pérez & Pacheco, 2006). Some laboratory experiments on Dickcissels captured in Venezuela tested the effectiveness of chemical repellents in reducing rice consumption, and found that both methiocarb and anthraquinone reduced consumption by 70% (Avery, Tillman & Laukert, 2001). Mitchell & Bruggers (1985) tested the effectiveness of methiocarb as a chemical repellent, as well as that of an olfactory (Tabebuia extract) and visual (blue carpenter’s chalk) deterrent, in reducing damage to cacao by woodpeckers in the Dominican Republic, but their results were inconclusive. Rodriguez et al. (1995) compared the effectiveness of methiocarb with that of a visual deterrent (calcium carbonate paint) in reducing eared dove damage to sunflowers and found that the latter was much more effective. Robles et al. (2003) found that using reflective objects as visual deterrents was more effective in reducing bird damage to quinoa than the chemical repellent bidrim. The effectiveness of a palatable repellent (Capsicum) and an olfactory repellent (Creolin) in reducing wildlife damage to corn in Colombia was tested, but no significant differences between the treatments and controls were found (Castillo-López et al., 2017). Thus, we believe that the use of non-lethal control techniques has been insufficiently tested and should be explored further to benefit the protection of biodiversity and the safety of commercial and subsistence crop plantations.

The results of our literature review point to a gap in the knowledge about vertebrate-crop conflicts in Latin America. However, it is important to highlight two aspects of the methodology we used that could bias our results. First, we did not include most kinds of gray literature in our review, and there may be more knowledge on the topic to be found in reports, or MSc and Ph.D. theses that are not published in scientific journals. However, conference proceedings were included in our review to minimize this bias (McAuley et al., 2000). Another shortcoming of our methodology is that we only performed searches using terms in English. While this is the main language used for scientific communication and publication, it is not the predominant language spoken in Latin America. We included studies returned by searches using terms in English but written in Spanish, Portuguese, or French. However, there could be more studies on this topic that would only be found by performing searches using terms in the languages spoken in Latin America. Despite these limitations, we believe that our review offers an accurate depiction of the published scientific literature on crop damage by vertebrates in Latin America.

Implications for management and future directions

Human-wildlife conflict is a pressing issue due to the simultaneous reduction of natural spaces and global human population growth. Crop damage is one of the most prominent reasons for conflict, as it affects the food security and economy of local communities. Despite this, research on crop-feeding by vertebrates in Central and South America is still emerging, and the body of literature on this topic is still limited. There is a lack of standardized methodologies to perform studies on crop damage, overreliance on farmers’ perceptions, and a lack of consensus on which protection techniques are preferable. As per our review, only 10 studies in the last four decades have performed experiments to test the effectiveness of crop protection techniques, seven of which tested non-lethal methods. Farmers tend to prefer lethal control methods that can endanger vertebrate populations, harm the environment, and negatively affect human health. We consider that there is an increased need to test non-lethal crop protection techniques in Latin America, which is already happening in Africa or Asia. Reliable and extensive experimentation should be conducted in different settings across Latin America to test which techniques work on different groups of vertebrates and crops that are involved in crop-feeding in the region.

Finding techniques that effectively protect crops from vertebrates without causing mortality is essential for solving this conflict while preserving both the environment and the interests of local communities. However, crop protection alone will not be able to solve this problem, and is only treating the symptoms and not the cause of the problem. Effective non-lethal protection methods need to be combined with a reduction in natural habitat loss and fragmentation, so that wild animals do not have to turn to agricultural products for food. Pairing effective non-lethal crop protection techniques with the conservation of native vegetation will reduce human-wildlife conflicts and help improve the quality of life of local communities while protecting native wildlife. Implementing such measures may present a difficult challenge owing to the current and predicted human population growth, current systems of production, and consumer-commodity model-based economies that demand increasing amounts of land for food production worldwide. Global agricultural development requires a systemic shift towards a more sustainable model that reduces the competition between food production and wildlife for land and resources.