Environmental factors affecting damage by wild boars (Sus scrofa) to rice fields in Yamanashi Prefecture, central Japan

Abstract

Wild boar Sus scrofa is widely distributed (Erkinaro et al. 1982; Saez-Royuela and Telleria 1986). Boars of this species often cause crop damage (Dardaillon 1986). Attempts to solve this problem have included nuisance control (McIlroy 1995), fencing (Hone and Atkinson 1983; Honda 2005), and feeding to entice wild boars away from crops (Mazzoni della Stella et al. 1995). However, the overall effectiveness of nuisance control in reducing crop damage has not been consistent. Some reports have described positive results (Conover and Decker 1991; Debernardi et al. 1995), but others have been negative (Mackin 1970; Andrzejewski and Jezierski 1978; Mazzoni della Stella et al. 1995). Similarly, fencing’s effectiveness for reducing crop damage has not been clarified (Positive report, Hone and Atkinson 1983; Honda 2005, Negative report, Geisser and Reyer 2004). Geisser and Reyer (2004) did not support supplemental feeding as a means to reduce crop damage. In other words, because no effective techniques exist, we are unable to cope effectively with crop damage without additional study. For example, a necessary study would identify, definitively, which environmental factors induce crop damage and predict damage. The importance of prediction is shown by forecasting of plant disease and pest which conducted in places throughout the Japan (Plant Protection Law in 2004 from the Ministry of Agriculture, Forestry and Fisheries). Additionally, Marsan et al. (1995) directly pointed out the importance of predicting wild boar damage to crops. Conover (2002) demonstrated the importance of manipulating the surrounding landscape for reducing damage. Before doing so, however, we must evaluate damage-related environmental factors and formulate a predictive model. The Meriggi and Sacchi (1992) model, which was estimated in Europe, was limited in its range of study area to a narrow area (20 km). Therefore, a follow-up study is needed. A shortage of necessary information pertains not only to damage prediction but also to estimation of the habitat selection of wild boars. This shortage of information is serious because habitat selection data are used as the basis for prediction. Wild boars are distributed from dry regions to tropical rainforests (Ralph and Maxwell 1985; Massei and Tonini 1992; Choquenot 1995): habitat selection should be estimated for each region. Regarding eastern Asia, only two studies related to habitat selection have been reported: Kodera et al. (2001) specifically examined vegetation and Park and Lee (2003) investigated topographic factors. However, no studies have covered both vegetationrelated and topographic factors. These two reports might show an erroneous conclusion because Kodera et al. (2001) neglected topographical factors and they could not exclude confounding (i.e. topographical) factors. As described above, environmental factors that contribute to damage have not been elucidated for eastern Asia; even the habitat selection of wild boars has not been closely investigated. In this study, we constructed a damage prediction model using a generalized linear model (GLM) and a geographical information system (GIS). To exclude confounding factors, all vegetation-related, topographical, and human disturbance factors were validated.