Does Land Management Intensity Influence Pollinator Assemblages and Plant–Pollinator Interactions in the Lowlands of Terceira Island (Azores)?

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study examined pollinator diversity, abundance, and composition in 3 land use types (natural, semi-natural pasture, and intensive pasture) on Terceira Island. They found that diversity and abundance did not differ across the three land use types, but that composition varied (although they did not conduct statistics to determine if differences were significant). The study is interesting and well-written and I enjoyed reading it. I have a few major and several minor comments that I hope will help strengthen the manuscript.

 

Major comments:

 

1. The authors set their study up as an examination of “grazing intensity,” but I wonder whether “land management intensity” is a more accurate description, given that the intensive pastures also re-seed their land yearly (reducing plant diversity), apply fertilizers, etc.

 

2. The NMDS plot shows differences between the three land use types visually, but I recommend also performing a PERMANOVA (function “adonis2”, package “vegan”), which can test whether or not the pollinator communities of the three land use types were significantly different.

 

3. What is your rationale for pooling the interaction data from all 10 sites per land use, rather than creating a separate network for each study site? If you create a separate network for each site, you will have 10 values per land use for each network metric, and you can then perform stats to see if any of the metrics are significantly different between land use types. I also think it’s more informative to know the number of plant species, pollinator species, etc. PER SITE for each land use, rather than the total number of plant species, pollinator species, etc. for each land use (see Table 2). In other words, I would like Table 2 to show the means and SE for each land use type for each network metric. Finally, when you pool data from different sites, you are combining species that did not have an opportunity to interact (e.g., pollinator species A is only found at site 1 and plant species B is only found at site 2), when in actuality they may have interacted if they were both present at the same site. This can inaccurately inflate network specialization.

 

4. This is just a suggestion, but it might be interesting to include pollinator species as a fixed factor in your GLMM examining pollinator abundance. Based on your networks in Figure 4 it looks like, while overall abundance did not differ by land use, the abundance of certain pollinator species did (e.g., Apis mellifera).

 

 

Minor comments:

 

L45-47: No reference cited. Please add supporting references. One suggestion is this paper that studied urbanization and agricultural intensification of bats, birds, and butterflies:

Olivier T, Thébault E, Elias M, Fontaine B, Fontaine C. 2020. Urbanization and agricultural intensification destabilize animal communities differently than diversity loss. Nature Communications, 11:2686.

 

L47-49: No reference cited. Here is one that examined native and exotic pollinator visits to exotic plant species:

Chitchak N, Hassa P, Traiperm P, Stewart AB. 2024. Who pollinates exotic plants? A global assessment across native and exotic ranges. Global Ecology and Conservation, 54:e03185​.

 

L64: Change to “threatened by”

 

L79: “territorial management” sounds odd. Do you mean “island land use management”?

 

L171: Approximately how much time was spent walking each transect?

 

L172: How did you determine whether a flying or resting insect was a pollinator? For example, I’m guessing that you recorded all bees as pollinators, regardless of whether they were flying, resting or visiting flowers. But what about flies/wasps/beetles/etc that were observed flying or resting, were they recorded as pollinators?

 

L189: Please briefly explain the distributional groups here so that readers have a general idea without having to search for Borges et al.

 

L193-200: It is unclear whether you calculated alpha diversity for plants or pollinators. Pollinators, yes? Please specify here so that it is clear to readers.

 

L432: Change to “assisted with”

 

Figure 3: In terms of the pollinator species that contributed most to differences between land use, only Xanthandrus azorensis seems to make sense. How do the other 4 species contribute to differences? What land use types were they associated with? In lines 281-282 you say that S. scipta and S. lunata occurred in most study sites, so how do they contribute to differences between land use?

 

Figure 4: Double-check all classifications of native/exotic. S. lunata is shown as orange (introduced) in the top network and green (native) in the other two networks.

 

Figure 4: Please include a large-size, high-resolution pdf version of this figure that includes ALL plant and pollinator species names in the SI.

 

Author Response

This study examined pollinator diversity, abundance, and composition in 3 land use types (natural, semi-natural pasture, and intensive pasture) on Terceira Island. They found that diversity and abundance did not differ across the three land use types, but that composition varied (although they did not conduct statistics to determine if differences were significant). The study is interesting and well-written and I enjoyed reading it. I have a few major and several minor comments that I hope will help strengthen the manuscript.

Major comments:

1. The authors set their study up as an examination of “grazing intensity,” but I wonder whether “land management intensity” is a more accurate description, given that the intensive pastures also re-seed their land yearly (reducing plant diversity), apply fertilizers, etc.

We agree with your suggestion and have made some changes throughout the manuscript acknowledging that the differences between study areas result from different land use management and have also slightly changed the title. However, we think that grazing intensity is the major driver of differences between study areas due to the combined effects of grazing, trampling and nutrient inputs from cow dung and their continuous effects year-round. For these reasons, we aimed to highlight this issue in the manuscript.

1. The NMDS plot shows differences between the three land use types visually, but I recommend also performing a PERMANOVA (function “adonis2”, package “vegan”), which can test whether or not the pollinator communities of the three land use types were significantly different.

Following your advice, we have now tested for significant differences on pollinator communities between land use types and present these findings. We followed a GLM framework which is more robust to overdispersion than PERMANOVA or ANOSIM using package mvabund (Wang et al. 2012, https://doi.org/10.1111/j.2041-210X.2012.00190.x).

1. What is your rationale for pooling the interaction data from all 10 sites per land use, rather than creating a separate network for each study site? If you create a separate network for each site, you will have 10 values per land use for each network metric, and you can then perform stats to see if any of the metrics are significantly different between land use types. I also think it’s more informative to know the number of plant species, pollinator species, etc. PER SITE for each land use, rather than the total number of plant species, pollinator species, etc. for each land use (see Table 2). In other words, I would like Table 2 to show the means and SE for each land use type for each network metric. Finally, when you pool data from different sites, you are combining species that did not have an opportunity to interact (e.g., pollinator species A is only found at site 1 and plant species B is only found at site 2), when in actuality they may have interacted if they were both present at the same site. This can inaccurately inflate network specialization.

You are right! Using the 10 sites as replicates to provide information on mean and SE values and using this information to calculate network metrics is more informative and the correct way to do it. We provide the new results in the revised version of the manuscript and also made the necessary changes you indicated us (in table 2).

1. This is just a suggestion, but it might be interesting to include pollinator species as a fixed factor in your GLMM examining pollinator abundance. Based on your networks in Figure 4 it looks like, while overall abundance did not differ by land use, the abundance of certain pollinator species did (e.g., Apis mellifera).

Thank you for this suggestion! We have now tested for significant differences in abundance between habitat-types for several pollinator species (Calliphora vicina, C. vomitoria, Sphaerophoria scripta, Stomorhina lunata, Xanthandrus azorensis - species with a significant influence in the NMDS pattern when using permutation tests (p<0.01), and also Apis mellifera and Bombus terrestris) using pollinator species as a fixed factor in the GLMM. The results are presented and discussed, particularly the issue of the spread of the exotic honeybee due to its potential negative impacts on island biodiversity.

 Minor comments:

 L45-47: No reference cited. Please add supporting references. One suggestion is this paper that studied urbanization and agricultural intensification of bats, birds, and butterflies:

Olivier T, Thébault E, Elias M, Fontaine B, Fontaine C. 2020. Urbanization and agricultural intensification destabilize animal communities differently than diversity loss. Nature Communications, 11:2686.

Done.

L47-49: No reference cited. Here is one that examined native and exotic pollinator visits to exotic plant species:

Chitchak N, Hassa P, Traiperm P, Stewart AB. 2024. Who pollinates exotic plants? A global assessment across native and exotic ranges. Global Ecology and Conservation, 54:e03185​.

Done.

L64: Change to “threatened by”

Done.

L79: “territorial management” sounds odd. Do you mean “island land use management”?

Yes, it sounds much better. Thank you.

L171: Approximately how much time was spent walking each transect?

We took approximately 20 minutes to complete each transect and have now added this information in the manuscript.

L172: How did you determine whether a flying or resting insect was a pollinator? For example, I’m guessing that you recorded all bees as pollinators, regardless of whether they were flying, resting or visiting flowers. But what about flies/wasps/beetles/etc that were observed flying or resting, were they recorded as pollinators?

In Azores, the number of flower visitors is very low since this archipelago is remote and relatively young. Many pollinator groups are absent from these islands and the ones that arrived are poor in species, when compared with mainland and the other Macaronesian archipelagos of Madeira and the Canaries. For example, in Terceira, there are only 8 butterfly species, 11 bees, 12 hoverflies, 5 wasps (Vespidae+Crabronidae) and very few floricolous beetles and other flies. Following our experience of over a decade working on Azorean pollinators, we already know these species and we are aware that this is not what happens in most places where species identification is difficult and studies usually target just one or a few pollinator groups.

L189: Please briefly explain the distributional groups here so that readers have a general idea without having to search for Borges et al.

Done.

L193-200: It is unclear whether you calculated alpha diversity for plants or pollinators. Pollinators, yes? Please specify here so that it is clear to readers.

We have now clarified that we are referring just to pollinators.

L432: Change to “assisted with”

We corrected this wording as we intended to state what we have been witnessing in the Azores for the past two decades.

Figure 3: In terms of the pollinator species that contributed most to differences between land use, only Xanthandrus azorensis seems to make sense. How do the other 4 species contribute to differences? What land use types were they associated with? In lines 281-282 you say that S. scipta and S. lunata occurred in most study sites, so how do they contribute to differences between land use?

The results we indicated were based on the species scores of the NMDS, comparing abundances between habitat-types using permutation tests. The species we highlighted were those with significant differences in abundance between habitat-types (P<0.01). However, following your advice, we have now tested the differences in pollinator species abundance between habitat-types using pollinator species as a fixed factor in the GLMM and discussed these findings.

Figure 4: Double-check all classifications of native/exotic. S. lunata is shown as orange (introduced) in the top network and green (native) in the other two networks.

You are right, thank you for noticing it. We have now checked all classifications and provide a new figure.

Figure 4: Please include a large-size, high-resolution pdf version of this figure that includes ALL plant and pollinator species names in the SI.

Done. We have now included the more detailed figures (with all species names) in the SI.

Reviewer 2 Report

Comments and Suggestions for Authors

I have reviewed the manuscript entitled “Land use influences pollinator assemblages and plant-pollinator interactions in the lowlands of Terceira island (Azores)”. The study addresses an important topic regarding the impact of grazing intensity on pollinator communities and plant-pollinator networks in an oceanic island context. The manuscript is generally well-written, methods are sound, and the findings are relevant for both ecological theory and conservation practice. However, I have several comments and suggestions that could further improve the manuscript. Major Comments: 1.The study was conducted only one year during summer (August 1st–September 10th). The authors need to indicate how long the entire flowering period is, whether the pollinator observation time covers the flowering period of the entire plant community, and how long the observation time is at each study site. Because pollinator communities and floral resources can vary at different days, month and years. I recommend the authors discuss this limitation and consider whether the conclusions might be different if sampling included spring or early summer. 2.The analysis of pollinator networks in this paper does not have a main line, and it is recommended that the authors pay attention to the changes of alien plants and pollinators in the network in the three land use intensities. 3.The authors mention that plant functional traits may influence pollinator visits, but no trait-based analysis was conducted. I suggest the author added some floral traits (e.g., flower shape, color, lower abundance) in the MS to better explain the observed patterns in network structure. 4.The authors should more explicitly address the selection of network metrics, such as Nestedness, generality. The author need to explain the reason for each parameter selection and provide the cites. 5.While the study focuses on interaction networks, it does not measure pollination function of flowering plants (e.g., seed set, seed number). I encourage the authors to discuss how the observed changes in networks might translate into functional outcomes. Minor Comments: L21: Grazing intensification is known to reduce species diversity. This is an unsuitable expression, many studies have shown that moderate disturbance increases species diversity. L28 by 41 pollinator species? L30 about generality? Is it plant, pollinator, or network level? L121: Please introduce the plant flowering period, the diversity of flowering plants and pollinators at the study sites. L176:How long and how did you observe each site? L299: The author built a total of 30 pollination networks, right? If so, the authors can analyze the structural changes of pollination networks under the three grazing gradients. Addition, the author can provide the original network data of the Excel version in the data sharing section. Figure 4 (networks) is informative but could be improved with better resolution or color differentiation for introduced vs. native species. Stomorhina lunata is introduced or native species?? Table 2:the author can explain the meaning of each parameter, e.g. specialisation and web asymmetry, generality

Author Response

I have reviewed the manuscript entitled “Land use influences pollinator assemblages and plant-pollinator interactions in the lowlands of Terceira island (Azores)”. The study addresses an important topic regarding the impact of grazing intensity on pollinator communities and plant-pollinator networks in an oceanic island context. The manuscript is generally well-written, methods are sound, and the findings are relevant for both ecological theory and conservation practice. However, I have several comments and suggestions that could further improve the manuscript.

Major Comments:

1.The study was conducted only one year during summer (August 1st–September 10th). The authors need to indicate how long the entire flowering period is, whether the pollinator observation time covers the flowering period of the entire plant community, and how long the observation time is at each study site. Because pollinator communities and floral resources can vary at different days, month and years. I recommend the authors discuss this limitation and consider whether the conclusions might be different if sampling included spring or early summer.

We have now added information discussing the limitations of this study, particularly the sampling being limited in time (lines 534-541). We sampled during summer because in Azores this is the period with the lower number of rainy days and when most pollinator species are active. During the other seasons, rains are quite frequent posing constraints for pollinator sampling and the observation of interactions. Following your advice, we have now included the observation time during transect sampling and information on plant flowering phenology.

2.The analysis of pollinator networks in this paper does not have a main line, and it is recommended that the authors pay attention to the changes of alien plants and pollinators in the network in the three land use intensities.

We improved the information on pollinator networks on the main text, checked plant and pollinator classifications and provided a new figure.

3.The authors mention that plant functional traits may influence pollinator visits, but no trait-based analysis was conducted. I suggest the author added some floral traits (e.g., flower shape, color, lower abundance) in the MS to better explain the observed patterns in network structure.

You are right, thank you for your suggestion! We have now added some information on the plant traits to support data interpretation and discussion.

4.The authors should more explicitly address the selection of network metrics, such as Nestedness, generality. The author need to explain the reason for each parameter selection and provide the cites.

We agree and have now included information on the selected metrics. Furthermore, we also listed some references on this topic, where readers can obtain more detailed information on these network metrics.

5.While the study focuses on interaction networks, it does not measure pollination function of flowering plants (e.g., seed set, seed number). I encourage the authors to discuss how the observed changes in networks might translate into functional outcomes.

We have now added information (lines 541-553) discussing the limitations of this study, including not having measured pollination effectiveness between study areas. Also, we comment on the potential consequences of land use intensification for the study ecosystems.

Minor Comments:

L21: Grazing intensification is known to reduce species diversity. This is an unsuitable expression, many studies have shown that moderate disturbance increases species diversity.

Yes, you are right. We have rephrased the sentence.

L28 by 41 pollinator species?

Yes, thank you.

L30 about generality? Is it plant, pollinator, or network level?

Generality is a network metric that measures the mean effective number of flowering plant species per pollinator species. We explain its meaning in line 347.

L121: Please introduce the plant flowering period, the diversity of flowering plants and pollinators at the study sites.

We added information on plant and pollinator diversity of the study area (lines 150-153) and the flowering period is reported in the following section.

L176: How long and how did you observe each site?

We took approximately 20 minutes to complete each transect. During transect walking we took note of the pollinators and plant-pollinator interactions. In case of doubt with the identification of a pollinator, the individual was collected with a sweeping net. Later, we walked the transect to count the number of flowers, identify all plant species in bloom and record vegetation height.

L299: The author built a total of 30 pollination networks, right? If so, the authors can analyze the structural changes of pollination networks under the three grazing gradients. Addition, the author can provide the original network data of the Excel version in the data sharing section.

Following the advice of another referee, we added new information on this topic and have also improved the analysis and changed figures and tables. Also, more detailed figures (with all species names) are made available as SI and the pollinator data (including interactions) was made publicly available in GBIF (http://ipt.gbif.pt/ipt/resource?r=pollinators_terceira)

Figure 4 (networks) is informative but could be improved with better resolution or color differentiation for introduced vs. native species. Stomorhina lunata is introduced or native species??

You are right, thank you for noticing the error with the colour coding of Stomorhina lunata. We have now checked all classifications and provide a new figure. We followed the colour codes for native and introduced species that we used in previous publications.

Table 2 the author can explain the meaning of each parameter, e.g. specialisation and web asymmetry, generality.

We now explain the selected network metrics in the main text and also include some references where more detailed information can be obtained.

Figures and tables must be improved

We made some changes on figures and tables, and will also follow the guidelines of the editorial board on this issue.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Many thanks to the authors for taking the time to thoroughly address my comments, I am glad they were helpful! I have one more minor comment, but other than that I think this manuscript looks much improved.

L234-235: When you say the 30 study sites were “group” it sounds like you pooled the data from all sites to create one network, or 1 network per land use type. I suggest that you use the following wording: “The plant-pollinator interaction data recorded at the 30 study sites were used to create bipartite networks using the bipartite package [62], with a separate network created for each study site.”

Author Response

Dear reviewer,

thank you very much for your helpful corrections, suggestions and comments which had a positive impact in improving our manuscript. Also, your kind suggestions to improve the English were most welcome.

Following your advice, we have changed the sentence of lines 234-235.

Best wishes

Reviewer 2 Report

Comments and Suggestions for Authors

I have no morecomments

Author Response

Dear reviewer,

thank you very much for your comments which helped us to improve our manuscript by providing more detail of the research work and also acknowledging the study limitations.

Best wishes