Efficiency of spreading maize in the garrigues to reduce wild boar (Sus scrofa) damage to Mediterranean vineyards
- First Online:
- Cite this article as:
- Calenge, C., Maillard, D., Fournier, P. et al. Eur J Wildl Res (2004) 50: 112. doi:10.1007/s10344-004-0047-y
- 270 Views
The aim of this work was to assess the efficiency of dissuasive spreading of maize to reduce the level of wild boar damage to vines in a Mediterranean area (Puéchabon, southern France). The 50 wine growers of Puéchabon were all questioned about the annual losses caused by the wild boar in each vineyard of the study area between 1990 and 1992. We also studied the distribution of the damage on a smaller scale, by mapping the damaged vines within two vineyards. In summer 1993, we distributed 4.7 tons of maize in the woods , and then questioned the wine growers about the losses in each vineyard for 1993. During 1990–1992, on both large and small scales, the damage had a patchy distribution, with more patches of damage occurring close to the woods. Moreover, the later the vines ripened, the less frequent the damage. The severity of the damage was only affected by the distance of the vineyard from woods. In total, the wild boar consumed 20,049 kg of grapes each year between 1990–1992 (193 kg/ha), and 63% of the vineyards were damaged. In 1993, both the density and the compensation paid in the département increased threefold. However, in our study area, both the proportion of damaged vineyards (36%) and the level of damage in the vineyards (151 kg/ha) were reduced, saving more than 60% in compensation. The dissuasive spreading of maize is therefore an efficient tool to reduce the level of damage to vineyards.
KeywordsCompensationDeterrentMediterranean habitatPatchy distribution Supplementary feeding
During the last decades, the wild boar bag has steadily increased in France (37,500 wild boar were shot in 1973, 200,000 in 1993, and 343,700 in 1998), and in certain regions, crop damage is becoming increasingly unacceptable to the farmers (Vassant 1996). In the Mediterranean region in particular, the wild boar population has risen sharply (from 4,000 boars culled in 1973 to 62,000 in 1997). Compensation for wild boar damage to vineyards, the main local crop, has also increased (from €110,125 in 1978 to €463,340 in 1997). Because of the process of rural decline, more and more fields are left uncultivated, allowing a deeper penetration of the boars in the cultivated fields. In addition, this problem can only worsen with the gradual shrinkage of French viticultural areas increasingly planted with high quality vines. The concern of wine growers is therefore growing year by year.
Three-quarters of the wild boar’s food intake is made up of two or three main items (Leranoz-Urtasum 1983; Fournier-Chambrillon et al. 1996). These staples always belong to the most abundant and most easily accessible food source for the period of the year. In the Mediterranean region, the abundant crop of acorns produced by the holly oaks (Quercus ilex) is the wild boar’s main source of food from September to June. However, acorns on the ground do not keep long enough for the animals to subsist on them from one year to the next (Fournier-Chambrillon et al. 1996), and they are no longer available in summer. During this period, other natural foods are also limited (Fournier-Chambrillon et al. 1994a). In late summer, grapes are the only edible material available in large quantities in our study area.
There is therefore a considerable need for efficient management tools to reduce the level of damage to crops. Numerous methods have been developed to effectively limit such damage. Methods based on the creation of acoustic, olfactive or gustative aversions have not given satisfactory long-term results (Vassant and Boisaubert 1984; Vassant 1994a). On the other hand, electric fences are an effective means to prevent access to crops for wild boars and may be recommended for any type of crops (Boisaubert et al. 1983; Vassant and Boisaubert 1984). However, continuous surveillance is required to prevent the deterioration of both fences and batteries. In addition, electric fences may be efficient to solve local problems, but their large-scale use is problematic.
The most effective method for reducing the level of damage on crops seems to be the dissuasive spreading of maize. Indeed, this method has proved to be an efficient means for the protection of cereal crops (oats, wheat, etc.) in several areas (Vassant and Breton 1986; Vassant et al. 1992; Vassant 1994a, 1994b; Geisser 1998). In other countries in Europe, this method is even regulated by law (e.g. in Wallonie, Belgium, by ministerial decree published 13 June 2003). On the other hand, the efficiency of dissuasive spreading of grain to protect the vineyards still has to be fully tested. In fact, we do not know whether maize spreading in the garrigue scrublands during the warmest period of the year will prevent wild boar from going to the vineyards to feed on the sweet, water-rich fruit.
A good understanding of the factors that determine both the occurrence and the severity of damage is an essential step toward the development of this technique in a Mediterranean habitat (Geisser 1998). According to the literature, the main factors affecting the level of damage to crops are the species cultivated (Mackin 1970; Andrzejewski and Jezierski 1978; Merriggi and Sacchi 1992; Genov et al. 1995; Onida et al. 1995; Geisser 1998), the distance of the crops from woods (Merriggi and Sacchi 1992; Genov et al. 1995; Onida et al. 1995), the ripening period of the crop (Vassant and Breton 1986; Vassant et al. 1992; Lavoisier et al. 1996), the density of the wild-boar population (Andrzejewski and Jezierski 1978; Vassant 1994a; Geisser 1998; Spitz and Lek 1999), and the availability of natural food in the woods (Mackin 1970; Andrzejewski and Jezierski 1978; Genov et al. 1995).
In this study, we investigated the spatial pattern of the damage on both a large scale (distribution of the damage over the study area) and a small scale (distribution within two vineyards), because many authors have stressed the fact that resource selection by the animals may differ on different spatial scales (Johnson 1980; Otis 1997, 1998). Our aim was to identify the factors affecting both the occurrence and the severity of wild boar damage to crops, paying special attention to the ripening period of the grapes and to the distance of the crops from woods. The efficiency of dissuasive spreading of maize to limit wild boar frequentation of the vineyards was then tested, and finally we estimated the financial benefits of this type of operation.
Materials and methods
The study area is located in southeastern France, in the Hérault département (French district covering 615,000 ha), 35 km northwest of Montpellier, in the Puéchabon commune (latitude: 43°40′N, longitude: 3°36′E). This commune, covering 3,125 ha, is situated at the junction of the large, wine-growing plain of Béziers and the garrigue scrublands of Montpellier. The plain stretches far out into the wooded mountains over a width of 2–5 km. The wild boars frequently cross this plain when they pass from one plateau to another, venturing into the vineyards to feed when the grapes are ripe (Fournier-Chambrillon et al. 1994b). Between 1990 and 1992, the hunting bag was 0.7 animals culled per km2. The hunting bag increased to 2 animals culled per km2 in 1993, because of a management policy aimed at increasing the wild boar population over the whole Hérault département (limitation of both the number of culled adult females and the duration of the hunting season).
The garrigue (covering 3,000 ha), made up of low woody plant communities of typical Mediterranean species (Phillirea latifolia L., Pistacia lentiscus L., Pistacia terebinthus L., Buxus sempervirens L., Juniperus oxycedrus L., Quercus coccifera L.), is interspersed by coppices of holly oak (Quercus ilex L.) with some Aleppo pines (Pinus halepensis Mill.) and pubescent oak (Quercus pubescens Willd.). Acorns produced by the holly oak are the main food source of the wild boar in fall and winter (Fournier-Chambrillon et al. 1994b). Maillard (1996) estimated that the mast crop on the commune varied roughly between 300–500 kg ha-1, depending on the tree density. The overall mast crop each year on the whole commune varies therefore between 900–1,500 tons. However, these acorns are no longer available in summer (most of them are eaten or rotten; Maillard 1996).
In our study area, the vineyard covers 125 ha, i.e. 45% of the farmed land, the rest being made up of fallows, fodder crops and olive groves which, for the greater part, have been left unattended. The crops found in the study area are representative of the region. Here, one finds the traditional varieties of vines that are used to make the wine from the hillsides of Languedoc. Within the study area, 283 vineyards were counted. Grapes are the most important crop, with very high yields. The different varieties of grape ripen at different times, which may also vary from year to year. The early varieties ripen between mid-July and mid-August and are not very common (2.3% of the total area planted with vines). Others ripen between mid-August and mid-September; they represent 34.4% of the vineyards. Finally, there are varieties which reach maturity between mid-September and mid-October. They are the most widespread and cover 63.3% of the vineyards.
Survey of wild boar damage to vineyards
Distribution of damage within the study area
To assess the impact of wild boar damage to the vineyards, a survey was carried out among all the wine growers in the study area, in 1992. The 50 vineyard owners were questioned regarding the degree (estimated in kilograms) of damage they had suffered in their vineyards from 1990–1992. According to the wine growers, the distribution of the damage in our study area did not change between 1990 and 1992, and the compensation paid was similar for these 3 years (between €6,000 and €6,300). Thus, they did not inform us on the level of damage to vineyards each year, but rather gave us an average annual estimation of the losses in each vineyard over this period. The level of damage in each vineyard was then mapped using Arcview GIS (ESRI 1996). The vineyards for which the estimated losses were higher than 300 kg ha−1 were considered to be severely damaged.
contiguous with the woods,
not contiguous with the woods but <50 m from the woods’ edge,
between 50 and 100 m,
between 100 and 400 m and
further than 400 m from the woods.
the “early varieties”, which ripen between mid-July and mid-August,
the “normal varieties”, which ripen between mid-August and mid-September, and
- 3.the “late varieties”, which ripen between mid-September and mid-October (Fig. 1).
We tested the main effects of distance from the woods and ripening period, as well as interactions between these factors, on the probability of damage occurrence using the generalized linear model, with a logit link (Mc Cullagh and Nelder 1989). We repeated the same analyses for the occurrence of severe damage, restricting them to damaged vineyards. However, due to the low number of damaged vineyards, we could not test the effect of interactions between distance from woods and ripening period on the frequency of severe damage. We therefore considered that these two factors were additive. We used the R software (Ihaka and Gentleman 1996) to perform these analyses.
Distribution of the damage within two vineyards
0% of the grapes consumed by wild boar,
>75% of the grapes lost.
Within each vineyard, we considered whether the frequency of wild boar damage was clumped in distribution, using a Geary test. We then used the lowess regression to smooth the damage distribution within the vineyard (Chessel et al. 1997). We performed this regression on the number of neighbours that minimised the lowess error. These analyses were carried out using the ADE-4 software (Thioulouse et al. 1997). Finally, we tested the correlation between the distance from woods and the level of damage using Spearman coefficients.
Dissuasive spreading of maize
In 1993, 4.7 tons of grain maize were distributed in the woods between 28 July and 6 September, forming a 4.5 km trail situated between 500 m and 1 km from the vineyards (an average amount of 25 kg day−1 km−1). The maize was poured in a steady sideways flow from a tractor-mounted tank. The daily distribution took on average 3 h. No other deterrent (e.g. electric fences) was applied in our study area during the focus period (1990–1993). To assess the efficiency of this technique to reduce the damage, we again questioned the 50 wine-growers in the study area on the damage (estimated in kilograms) suffered that year. For each vineyard, we determined whether the damage in 1993 was reduced, increased or similar to previous years. We then modelled the probability that a vineyard was less damaged during the application of the deterrence method using a generalised linear model with a logit link, restricting these analyses to the vineyards damaged during previous years. As for the models of severe damage, we considered that the effects of distance from woods and time of ripening of the vines were additive. We used the R software (Ihaka and Gentleman 1996) to perform these analyses.
Survey of wild boar damage to vineyards
Distribution of damage within the study area
The frequency of damage was also affected by the ripening period of the varieties cultivated (χ2=15, df=2, P<0.001). The earlier a vine reached maturity, the higher the frequency of damage. On the other hand, we did not identify any effect of the ripening period on the frequency of severe damage (χ2=3.4, df=2, P=0.18). We found no effect of the interactions between the ripening period and the distance from the woods on the probability of damage (χ2=9.9, df=7, P=0.19).
Distribution of the damage within two vineyards
Efficiency of dissuasive spreading of maize to reduce damage
Comparison of the level of damage to vineyards without (1990–1992) and with (1993) dissuasive spreading of maize in the vineyards of Puéchabon (south of France)
Total loss (kg)
No. of damaged vineyards
Compensation paid (€)
In 1990–1992, both the proportion of vineyards damaged by wild boar and the intensity of the damage increased with the proximity to woods. We also identified a clumped distribution of the damage. In addition, varieties of vine that ripen in early summer were more frequently damaged than others. The dissuasive spreading of grain maize resulted in a slight decrease in both the proportion of damaged vineyards and the intensity of the damage in our study area.
Wild boar preferentially wander along known paths for foraging, causing damage in the neighbourhood of selected foraging places, resulting in a patchy distribution of the damage (Kristiansson 1985; Welander 2000). Thus, even if the losses were rather low on the scale of the study area (3.5% of the grape production), locally, certain vineyards were strongly damaged, and certain wine growers were more affected by the losses than others. On both large and small scales, the patches of damage occurred mainly close to the woods. Security seems to be a factor of the utmost importance for the wild boar (Spitz and Janeau 1995). The animals will use the vineyards while taking care to stay close to the field edges, allowing easier escape in case of threat. This kind of behaviour has also been noted in other areas (Vassant and Breton 1986; Gérard and Campan 1988; Merriggi and Sacchi 1992; Genov et al. 1995; Onida et al. 1995; Spitz and Lek 1999).
As long as sufficient food is available in the woods, the wild boar scarcely feed on crops (Andrzejewski and Jezierski 1978; Mackin 1970; Genov et al. 1995). In our study area, there is a constant increase in the availability of food in the woods from July to October (Fournier-Chambrillon et al. 1996). The summer (July–August) is a period of food shortage, and the wild boar feed frequently on vines. From September, there is a gradual increase in acorn availability, and consequently, a decrease in the frequency of the damage to vines. Thus, the later a vine ripens, the less frequent the damage.
The dissuasive spreading of maize results in an increase of the food availability in the woods, and many authors have stressed its efficiency to reduce the damage to cereal crops (Vassant and Breton 1986; Vassant et al. 1992; Vassant 1994a, 1994b; Geisser 1998). This method also seems efficient to reduce damage to vines. In our study area, the hunting bag increased threefold between 1990-1992 (wild boar culled, 0.7/km2 ) and 1993 (wild boar culled, 2/km2) in the woods of Puéchabon, whereas the hunting pressure did not change between the two periods (similar number of hunters, number of days hunted per year and numbers of hounds), suggesting an increase in the wild boar population in 1993. This augmentation was also recorded for the whole Hérault département (from 3,500 to 5,500 wild boar culled), which led to a considerable increase in the level of damage to crops: compensation paid yearly on the whole département increased from €24,400 in 1990–1992 to €544 000 in 1993. However, in our study area, we observed a slight decrease in both the level of damage and the number of damaged vineyards in 1993. Although we acknowledge that a study of the evolution of the availability of natural food in the woods between 1990–1992 and 1993 would have strengthened any conclusion, this decrease is certainly the consequence of the dissuasive spreading of grain. This therefore emphasizes the efficiency of this measure of protection.
The fact that the dissuasive spreading of maize lead to a decrease in the level of damage was expected. Actually, we need an estimation of the financial gain of such a management tool to inform land managers on the efficiency of this method. The current cost of maize spreading may be estimated at €3,461 (4.7 tons of grain maize: €1,433, equipment for the distribution of maize: €381, running expenses for a vehicle for 40 days × 15 km: €183, wages for 40 days × 3 h: €1,464).
In 1993, when the deterrence measure was applied, damage compensation paid amounted to €4,797. To draw up a complete economic balance-sheet, we also estimated the cost of damage compensation that would have been paid in our study area in the absence of deterrence measures. In fact, in 1993, damage to the vineyards in the département increased 3.4-fold compared to the three preceding years. A similar trend was recorded in various sectors of the same département surrounding our study area. So, to reach an estimate of the expenses which, according to our study, should have been incurred, we multiplied the expenses for wild boar damage in the study area in the years 1990–1992 (estimated at €6203) by 3.4, i.e. a total of €21,090.
Estimated losses without supplementary food: €21,090
Estimated losses with supplementary food: €4,797
Cost of dissuasion: €3,461
Net final balance: €12,832
In addition to the financial gains, a further positive aspect is that the amount of damage went down. This satisfies the farmers and thus minimises sources of conflict between farmers and hunters.
After this study, dissuasive grain-spreading has been continued by the hunters in Puéchabon, also around some isolated vineyards in the forested hill sides. Every year, the people concerned meet in June, and devise a plan of action for the next protection campaign based on the mast crop and last year’s hunting bag, on the understanding that the hunters are responsible for the daily distribution of the maize and the erection of crop fences and the farmers ensure their upkeep. This action was completed by the construction of a few electric fences to protect the most sensitive vineyards, i.e. the vineyards planted with early variety of vines and close to the woods. In spite of a constant increase in wild boar numbers, their pressure on the vineyards is negligible and a climate of good understanding has been reinstated between wine growers and hunters (Maillard and Sanier 1996).
We emphasise that maize should not be provided at localised points. Rather, supplying the maize over 5–10 m wide strips guarantees that the supplementary food is accessible to all wild boar and that one male or a group of females does not drive the other wild boar off (Vassant et al. 1992). The maize should be supplied as long as the crops are ripe, at a rate of 2–4 kg/100 m, according to the recommendations of Vassant et al. (1992). The managers should acknowledge that the dissuasive spreading of maize is just a palliative in the protection of crops and is unlikely to be as efficient in cases of overdense population. Thus, this measure should be added to other measures intending to control the population density, as part of an effective policy of management of wild boar populations.
We are grateful to the wine growers of Puéchabon who answered our questions, as well as to the many students who helped during data collection. We also warmly thank Eric Baubet (Office national de la chasse et de la faune sauvage) for his useful comments on an earlier draft of this paper. The experiments carried out in this study comply with the current laws of France.