Mutualistic networks provide insights into large ecological questions about global patterns of community structure, but they can also supply practical information in a local, applied setting. Plant-pollinator networks can (i) provide insights into indirect interactions affecting species of conservation concern, (ii) inform restoration practice by identifying plants playing key pollinator-supporting roles, and (iii) provide information about how and whether plant-focused restoration is bringing back wider community complexity. We used two years of observational networks to examine pollination interactions of a pollinator-dependent federally threatened plant (Castilleja levisecta) and a pollinator-dependent host plant for endangered butterflies (Castilleja hispida), in six sites along a gradient of restoration. Sites ranged from 0-12 years of restoration via prescribed fire, herbicide, and seeding. Visiting pollinators were captured at patches of every flowering forb species in 30-minute observations throughout the season. We examined the resulting networks for: indirect interactions of Castillejas; identities of plant “hubs” supporting >20 species of visiting insects; and community metrics such as nestedness and modularity.
Results/Conclusions
(i) Conservation application. We observed only two pollinator species visiting C. hispida and six visiting C. levisecta, and all were shared with other native and non-native plants in the networks. At some sites, non-native plants were the only observed providers of floral resources to Castilleja pollinators after Castilleja flowering had ended; these non-natives may act as indirect facilitators of Castilleja seedset by supporting pollinator populations. Best management practices should therefore include removing non-native forbs in stages rather than attempting to eliminate them wholesale, risking a gap in resources for pollinators of key native plants. (ii) Restoration practice application. Six native forbs and three non-native forbs supported >20 pollinator species. Native “hub” plant species, planted as early as possible in the long trajectory of restoration, could support pollinators that in turn visit other native plants when they are seeded or planted later on. Non-native “hub” plants support pollinators, yet are often targeted for removal; if removed, they should be replaced with phenologically and nutritionally comparable native plants. (iii) Tracking community change with restoration. Network metrics such as nestedness, connectance, complementarity, and modularity changed systematically with restoration, and documented indirect changes in the wider plant-pollinator community as plant composition was altered. As catalogs of indirect interactions, networks provided simple and direct insights that could inform conservation and restoration actions, as well as information about restoration-influenced changes in community structure.