Influence of landscape structure on endemic cicadas in New Zealand kiwifruit orchards

Abstract

1 The endemic cicada species Amphipsalta cingulata (Fabricius) and Amphipsalta zelandica (Boisduval) are pests of New Zealand kiwifruit. 2 We determined the abundance of A. cingulata and A. zelandica by counting finalinstar exuviae in a block of 'Hayward' kiwifruit, the dominant cultivar, on each of 70 blocks on separate orchards in the Bay of Plenty, New Zealand. 3 We used a geographic information system and fragstats to generate predictive variables describing landscape structure in four nested landscapes ranging in size between 6.25 and 400 ha for each site. Other variables described the physical characteristics of the site and management practices. Data were analyzed by boosted regression trees, a method that combines the advantages of regression trees and machine learning. 4 The most influential variables differed for each species. Modified coastal landscapes with high densities of 'Hayward' kiwifruit were most favourable for A. cingulata. For A. zelandica, favourable landscapes contained significant areas of native forest.The 12 most influential variables accounted for 51% and 46% of the total influence of all variables measured for A. cingulata and A. zelandica, respectively. 5 Landscape structure was more influential than insecticide use and local site factors.Despite the apparent low vagility of cicadas, landscape structure at relatively large scales of ≥25 ha was influential for both A. cingulata and A. zelandica. The ability to use a wide range of hosts within the production landscape may account for this pattern. Key variables need to be confirmed by identifying the same patterns in other landscapes.

Influence of the surrounding landscape on crop colonization by a polyphagous insect pest

“…The relative importance of particular habitat patches on pest population dynamics has been investigated in focal‐patch landscape studies (Brennan et al., ), where correlative methods have been used to quantify the influence of particular patches on the relative abundance of a focal species (Pope et al., ; Carrière et al., , , ; Ricci et al., ; Bahlai et al., ; Logan et al., ). The data sets used in these studies, however, are often modest in size, due to the high sampling effort required to estimate the density of a focal species in a large number of fields and to characterize the landscape composition.…”