4.3 Different effects of nonforest landscapes on gene flow and dispersal
Urban landscapes are generally considered to be barriers to gene flow, and an increase in farmland landscapes can also directly or indirectly negatively affect the ecological processes and dispersal of many species (Butler, Vickery, & Norris, 2007; Johnson & Munshi-South, 2017; Lane-Degraaf, Fuentes, & Hollocher, 2014; Ruell et al., 2012; ). According to the Mantel test of the correlation between the LCP of multiple resistance surfaces and the genetic distance, they were significantly correlated only when the urban and farmland landscapes were given high resistance values (Figures 4A, B). By contrast, no significant correlations were obtained with the other three resistance surfaces in which low resistance was given to urban and farmland landscapes (Supporting information Figure S3). Thus, by acting as barriers, urban and farmland landscapes played an important role in the dispersal and the gene flow of M. alternatus . This effect is possibly due to the limited food sources of the species in cities and farmland. Moreover, the use of pesticides in farmlands can also affect insects (Sonoda, Izumi, Kohara, Koshiyama, & Yoshida, 2011), likely affecting the identification of hosts by M. alternatus and inhibiting dispersal.
Notably, the LCTA, in which the resistance value was objectively defined, also indicated that urban and farmland landscapes had a negative effect on gene flow (Tabel 3). Especially in Xiapu, the best three models in the LCTA showed that urban landscapes inhibited gene flow. In addition, the GAM model also suggested that urban and farmland landscapes were negatively correlated with genetic diversity at the 800 to 1,000-m scale (Figure 6). Urbanization is one of the important factors that cause the loss of species habitat as well as forest fragmentation and the gradual but eventual disappearance of small areas of forest (Caspersen & Olafsson, 2010), and the distribution of farmland destroys the continuity of host distribution. Because of the decline in forest coverage and the disappearance of host continuity, the population of M. alternatus may be separated to form isolated populations, thus reducing the gene flow. Indeed, farmland can directly affect the distribution and spread of insects and even hinder insect migration, which has important effects on population dynamics (Ge, Ou, & Men, 2016). In a study on other species of Monochamus , the flying time of this species was short, sometimes only 1 km (David, 2014), so the farmland and urban landscapes at the scale of 800 to 1,000 m would most likely be barriers due to the unwillingness to fly. Thus, at a fine-scale, urban and farmland landscapes distributed are important barriers to dispersal and migration in this species.
Corridors in landscapes can promote the dispersal of species among different landscape types (Lapolla, 1993; Nicholls, Parrella, & Altieri, 2001). Among nonforest landscapes, roads are very important corridors, and especially in areas with high landscape diversity, road patches can often play a role in connectivity. The fourth model of the LCTA in Shunchang showed that roads were the most suitable dispersal habitat for this species (Table 3). The LCP model showed that roads had the same low resistance value as other host landscapes (Figure 4C). The db-RDA and GAM models also indicated that roads were positively correlated with genetic diversity (Figure 5, 6;Table 4, 5). These results all suggest that roads are important and promote the dispersal of M. alternatus in a heterogeneous landscape. Although the existence of road networks can inhibit the dispersal and gene flow in some other species (Garcia-Gonzalez, Campo, Pola, & Garcia-Vazquez, 2012; Hartmann, Steyer, Kraus, Segelbacher, & Nowak, 2013), M. alternatus has the characteristic of the human-mediated movement, such as in the transport of infested wood from one area to another, including lumber and wood packaging material (Haack, 2006; Ye, 2019). This type of movement may cause the dispersal of this species in heterogeneous landscapes and may even lead to long-distance transmission. In a previous study on the genetic structure of M. alternatus , the road transportation system was also found to promote the dispersal of this species in mainland China (Hu, Ning, Fu, & Haack, 2013). Therefore, roads can facilitate dispersal and gene flow of M. alternatus in heterogeneous landscapes by promoting long-distance, human-mediated transport.