Species traits and network structure predict the success and impacts of pollinator invasions

Species invasions constitute a major and poorly understood threat to plant–pollinator systems. General theory predicting which factors drive species invasion success and subsequent effects on native ecosystems is particularly lacking. We address this problem using a consumer–resource model of adaptive behavior and population dynamics to evaluate the invasion success of alien pollinators into plant–pollinator networks and their impact on native species. We introduce pollinator species with different foraging traits into network models with different levels of species richness, connectance, and nestedness. Among 31 factors tested, including network and alien properties, we find that aliens with high foraging efficiency are the most successful invaders. Networks exhibiting high alien–native diet overlap, fraction of alien-visited plant species, most-generalist plant connectivity, and number of specialist pollinator species are the most impacted by invaders. Our results mimic several disparate observations conducted in the field and potentially elucidate the mechanisms responsible for their variability.

The results are really interesting pointing at the combined importance of traits of the alien pollinator species and properties of the invading network (including trait space of the native pollinator species); for example alien pollinators with a certain trait (high foraging e fficiency) are always successful in invading the network, but their impact on the native community is dependent of the properties of the network. In order to disentangle the effects of the different parameters the authors use a C lassification Tree analyses. My question is though if it would not be more interesting to go into more depth with the statistical analyses, in particular looking for interactions among different terms. The interactions between network properties and traits of the alien pollinator wo uld be of particular interest. One suggestion is to use a model selection approach and, due to the large number of parameters to fit, build up from very simple to more complex models that then would include interactions. I think this would indeed strengthe n the message of this paper. (One additional effect of identifying potential interacting effects would also be that it would allow the authors to elaborate more in their text on the importance of network structure which could be valuable.) My second comment is regarding the consumer-resource-model used. This model requires that numerous parameters are described both for the resources and the pollinators. I know the authors in earlier papers preformed sensitivity analyses and was slightly surprised this wa s not included here. It might be that those analyses could be used here as well, but if so at the very least I think the authors should address this in the method section in the main text.
Other than that, I really enjoyed reading the paper and think it would be of interest for a wide audience.
Reviewer #2 (Remarks to the Author): The authors establish a mathematical model to study the impact of in vading pollinator species in plant pollinator networks. Using extensive numerical simulations over large ensembles of networks of different size and connectance, they study the influence of pollinator foraging traits and network structure for invasion success and impact on native communities.
The study is technically performed at a high level, it includes impressive numerical simulations and a comprehensive analysis of all possible cofactors influencing the numerical outcomes. The methods and results are clearly presented. Thus, it is a study that deserves being published in a major ecological journal.
I have second thoughts, however, if this study is suitable for Nature C ommunications. Even though the model setup and the results are undoubtably novel, most findings are not really astonishing. To give some examples, major findings such as "invaders impact native pollinators primarily by strongly decreasing the floral rewards of the plant species that invaders visit", "sharing of links between a native pollinator species and the invader reduces the floral resources", "these effect becomes stronger in networks whose most generalized plant species interacts with most of the pollinators in the network", etc. would be the first ideas that come to mind even without numerical simulations. That is not to say that I find the study superfluous. In contrast, the differentiated and quantitative analysis of the many factors influencing invasion success and impact definitely has many merits and provides many opportunities for future studies. But, I am afraid that this paper does not provide a real breakthrough and will have only a minor influence on thinking in the field. What could have brought this paper over the top would be a strong connection of simulation results t o empirical data (even though I can understand that this may not (yet) be possible for several reasons).
In summary, I recommend to resubmit this paper to a strong ecological journal.
Response to the referees' comments on manuscript NCOMMS-18-03578-T Please see below our point-by-point response to each reviewer's comment. Our responses are in italic-bold font here and the edits to the manuscript are yellow-highlighted in the main text.
Reviewer #1 (Remarks to the Author): This paper addresses the very interesting and timely question of which factors are driving the invasion success of alien pollinator species in complex resource -pollinator networks. The authors simulate the pollinatorresource dynamics over time using a dynamical model developed a few years ago by the authors themselves. The main results are that foraging efficiency of the alien pollinator predict their invasion success but that the impact the invasion has on the native community is dependent on the structural properties of the network. The paper is clearly written, addresses and interesting topic and also have an important message.

We thank the reviewer for the interest and time spent on our manuscript, and for the positive comments on our main results, narrative and relevance of our study.
However, I have two comments/suggestions that I really think would improve and strengthen the paper.

We thank the reviewer for these excellent suggestions that greatly improved our manuscript. We believe have thoroughly addressed them.
The results are really interesting pointing at the combined importance of traits of the alien pollinator species and properties of the invading network (including trait space of the native pollinator species); for example alien pollinators with a certain trait (high foraging efficiency) are always successful in invading the network, but their impact on the native community is dependent of the properties of the network. In order to disentangle the effects of the different parameters the authors use a Classification Tree analyses. My question is though if it would not be more interesting to go into more depth with the statistical analyses, in particular looking for interactions among different terms. The interactions between network properties and traits of the alien pollinator would be of particular interest. One suggestion is to use a model selection approach and, due to the large number of parameters to fit, build up from very simple to more complex models that then would include interactions. I think this would indeed strengthen the message of this paper. (One additional effect of identifying potential interacting effects would also be that it would allow the authors to elaborate more in their text on the importance of network structure which could be valuable.) Table S4 in the Supplementary Information for the full models. Fig 3 (see added  text L160-162). Fig. 1

depicting Outlier Box Plots.
My second comment is regarding the consumer-resource-model used. This model requires that numerous parameters are described both for the resources and the pollinators. I know the authors in earlier papers preformed sensitivity analyses and was slightly surprised this was not included here. It might be that those analyses could be used here as well, but if so at the very least I think the authors should address this in the method section in the main text.

We thank the reviewer for pointing out this important subject. The revised version of our methods (L290-294) now includes: "Sensitivity analyses of the dynamic model have been performed in previous studies34,39, and the main results presented here (i.e. foraging efficiency of alien pollinators is sufficient to predict invasion success, while information on the network structure is also required to predict the invaders' impact on natives) are qualitatively robust to variation in parameter values."
Other than that, I really enjoyed reading the paper and think it would be of interest for a wide audience.