The effect of elevation, latitude, and plant richness on robustness of pollination networks at a global scale

Plant-pollinator interactions play a vital role in the maintenance of biodiversity and ecosystem function. Geographical variation in environmental factors can influence the diversity of pollinators and thus, affect the structure of pollination networks. Given the current global climate change, understanding the variation of pollination network structure along environmental gradients is vital to predict how global change will affect the ecological interaction processes. Here, we used a global plant-pollinator interaction data collection by the same sampling method at the same period to explore the effects of elevation, latitude, and plant richness on the structure and robustness of pollination networks. We analyzed a total of 87 networks of plant-pollinator interactions on 47 sites from 14 countries. We conducted a piecewise structural equation model to examine the direct and indirect effects of elevation, latitude, and plant richness on the network robustness and analyzed the function of network structure in elucidating the relationship between robustness and these gradients. We found that plant richness had both positive effects on robustness under random and specialist-first scenarios. Elevation, latitude, and plant richness affected network connectance and modularity, and ultimately affected network robustness which were mediated by nestedness under specialist-first and random scenarios, and by connectance under the generalist-first scenario. This study reveals the indirect effects of elevation, latitude, and plant richness on pollination network robustness were mediated by nestedness or connectance depended on the order of species extinctions, implying that communities with different pollination network structures can resist different extinction scenarios.