Could Pollination Networks be Applied for Biological Conservation and Monitoring of Threatened Plant Communities? – Examples from an Endangered Ecological Community, the Howell Shrublands

Abstract

<p>Plant-pollinator interactions play a pivotal role in the structure and persistence of biological communities. Despite this, plant-pollinator relationships are rarely considered in applied conservation, potentially resulting in counterproductive or ineffective management practices. By allowing interactions between plants and pollinators to be quantitatively assessed, network analyses may offer valuable information for applied conservation. To demonstrate the potential utility of network analyses in conservation, this study will attempt to inform management for an endangered ecological community in eastern Australia using a multi-year pollination network. Over six seasons, 1,633.4 hours of floral visitation data were collected for 103 plant species to construct a network of plant-pollinator interactions. Asymptotic richness estimates were used to evaluate sampling efficacy, and predict the amount of effort required to complete sampling. Network analyses were conducted using the Bipartite package for RStudio, allowing species of high conservation value to be identified in terms of interaction strength, and the pollinator service index (<i>PSI</i>). Network specialisation (<i>H₂</i>’), nestedness (<i>wNODF</i>), robustness, and modularity were assessed as potential indicators of community stability. Sampling was estimated to detect up to 90.05% of floral visitors, and 66.66% of interactions, with floral species predicted to require an average of 18.03 hours observation to sample a majority of interactions. This was also found to coincide with stable asymptotic richness estimates for a separate sample of species subject to relatively intensive sampling efforts and is therefore recommended as a baseline for future testing and sampling. Exotic honeybees (<i>Apis mellifera</i>) were identified as an important pollinator within the network, which was generally more specialised, modular, and robust to floral extinctions than could be explained by null models. Dividing the network into temporal sub-networks revealed late winter and early summer flowering plants were significantly less robust to pollinator extinction in spite of high interaction nestedness. This level of risk is likely to increase significantly if honeybees become locally extinct due to the parasite <i>Varroa destructor</i>. Therefore, conservation efforts should focus on late winter pollinators as a means of preserving the floral community.</p>