Network analysis for co‐occurrence of pest insects on host crops

Abstract

AbstractIn this study, the co‐occurrence patterns of 618 pest insects for 47 host crops, including vegetables, grains, and fruits, were identified. To identify the pest co‐occurrence patterns for various crops, and the interactions among the pest insects and crops, we employed social network analysis methods. We used three traditional centrality measures (degree, closeness, and eigenvector) to determine the relative significance of each crop and pest as a node in the network. Throughout the network analysis, crops and pest nodes were divided into six groups, based on modularity. Crops in the same group could be considered as alternate hosts for pests from the same group. There were clear differences in the cultural practices between groups (i.e., dryland farming versus wetland farming). This indicated that dryland crop pests do not use wetland crops as resources. Pome fruit trees, such as apple and pear, had high centrality indices, which indicated the importance of these crops in the network and their high vulnerability to damage by a multitude of pests. In this study, although it was assumed that all crops were cultivated on the same piece of land during a single growing season, the complex interactions between the whole units were visualized and analyzed as a computable network.