Evaluating the taxa that provide shared pollination services across multiple crops and regions

Abstract

Many pollinator species visit multiple crops in multiple regions, yet we know little about their pollination service provisioning at local and regional scales. We investigated the floral visitors (n = 13,200), their effectiveness (n = 1718 single visits) and response to landscape composition across three crops avocado, mango and macadamia within a single growing region (1 year), a single crop (3 years) and across different growing regions in multiple years. In total, eight wild visitor groups were shared across all three crops. The network was dominated by three pollinators, two bees (Apis mellifera and Tetragonula spp.) and a fly, Stomorhina discolor. The visitation network for the three crops was relatively generalised but with the addition of pollen deposition data, specialisation increased. Sixteen managed and wild taxa were consistently present across three years in avocado, yet their contribution to annual network structure varied. Node specialisation (d') analyses indicated many individual orchard sites across each of the networks were significantly more specialised compared to that predicted by null models, suggesting the presence of site-specific factors driving these patterns. Identifying the taxa shared across multiple crops, regions and years will facilitate the development of specific pollinator management strategies to optimize crop pollination services in horticultural systems.

Figure 1

Canonical correspondence analysis plot showing the distribution of pollinator taxa and orchard sites as a function of crop type (avocado, macadamia and mango) in Bundaberg, 2016. Grey arrows show significant explanatory factors with length of arrow indicative of strength of significance.

Figure 2

Insect visitor networks of mango, avocado and macadamia orchard sites in Bundaberg during flowering in 2016; (a) visitation network and (b) pollinator effectiveness network. Pollinator species shared between crops shown in colour and coded to different taxa. Height of each visitor group node reflects proportional abundance.

Figure 3

Family level visitation networks of avocado orchard sites in (a) Sunraysia region [15 orchard sites], and (b) Bundaberg [7 orchard sites] during flowering in 2015. Black nodes indicate orchard sites in each region. Visitor group labels colour coded to indicate shared groups (blue), unique groups (black) and unidentified visitors grouped to order (grey).

Figure 4

Visual representations of bipartite networks, highlighting different node types used in this study. Question 1 networks; (a) visitation network with crop type and insect visitor group as nodes; (b) pollinator effectiveness network with crop type and insect visitor group (visitation rate x mean pollen deposition) as nodes. Questions 2 and 4 networks; (c) regional visitation networks with individual orchard sites and taxonomic families as nodes; (d) visitation networks with individual orchard sites (unique crop, region and year) and insect visitor group as nodes. Question 3 networks.