Year-round temporal stability of a tropical, urban plant-pollinator network.

Plant-pollinator interactions are known to vary across time, both in terms of species composition and the associations between partner species. However, less is known about tropical pollination networks, and tropical urban parks provide a unique opportunity to study network stability in an environment where temperature and floral resources are relatively constant due to both the tropical climate as well as park horticulture. The objectives of this study were thus to examine the interactions between flowering plants and their potential pollinators in a large, tropical city (Bangkok, Thailand) across 12 consecutive months, and to assess the stability of network properties over time. We conducted monthly pollinator observations at 9 parks spaced throughout the city, and collected data on temperature, precipitation, floral abundance and floral species richness. We found that neither pollinator abundance nor richness varied significantly across months when all parks were pooled. However, pollinator abundance was significantly influenced by floral abundance, floral richness, and their interaction, and pollinator richness was significantly influenced by floral richness and precipitation. Finally, we found that network properties did not change across months, even as species composition did. We conclude that the year-round constancy of floral resources and climate conditions appear to create a network in dynamic equilibrium, where plant and pollinator species compositions change, but network properties remain stable. The results of this study provide useful information about how tropical pollinators respond to urban environments, which is particularly relevant given that most urban development is predicted to occur in the tropics.

Introduction

While the importance of plant-pollinator interactions has long been recognized [1,2], we have only recently developed the computational methods needed to analyze entire networks, and over extended periods rather than at a single point in time [3,4]. Analyzing complete pollination networks provides a more accurate understanding of the community, since both plants and pollinators typically interact with more than one partner [5-7]. Moreover, studying changes in pollination networks over time can (1) improve our understanding of pollination services, as most angiosperm species rely on animal pollinators [8,9]; (2) provide clarity on whether species are specialists, generalists, or sequential specialists [10,11]; (3) facilitate predictions of how plants, pollinators, and their interactions will respond to climate change [12]; and (4) provide critical information that can help with conservation efforts of both pollinators, which are undergoing population declines worldwide, and the plants that depend on them for reproduction [13,14].

Previous studies examining pollination networks over time have found a range of temporal patterns, from networks that appear highly stable to those that are highly dynamic [4]. Yet most work to date has been conducted in temperate or arctic regions [10,15-24] and we still know very little about the stability of tropical pollination networks (but see [25,26]). Thus, while ecologists are interested in finding universal patterns and mechanisms behind plant-animal networks [3], we still lack data for tropical systems.

Additionally, given the rate at which humans are changing the environment, it is important to study interactions not only in natural habitats, but also in human-modified landscapes [27]. For example, global urban land area is predicted to triple in size by 2030, compared to 2000, and most urban expansion is projected to occur in tropical areas [28]. Therefore, studying plant-pollinator interactions in cities can provide valuable information about how pollinators fare in one of the few habitat types that is expected to increase in size in the near future [27,28]. Finally, the abundance of floral resources found year-round in tropical cities [29], often due to the mix of native and exotic plant species found in cultivated parks, provides a unique opportunity to study plant-pollinator interactions in a relatively stable environment. Therefore, the objectives of this study were to examine the interactions between flowering plants and their potential pollinators at public parks in Bangkok, Thailand across 12 consecutive months, and to assess the stability of network properties throughout this period.

Methods

Study area

Data were collected in Bangkok, Thailand between December 2017—November 2018. Bangkok is the most populated city in Thailand (over 9.6 million residents according to the 2010 census) with very little natural habitat; nearly all vegetation is cultivated and managed [30]. The climate is tropical; average monthly temperatures range between 22–32°C and average annual rainfall in the area ranges from 1,100–1,600 mm [31]. We collected data from 9 parks spaced throughout the city (S1 Fig). These nine parks were selected because they are all at least 1 km apart, and represent a diversity of park sizes (S1 Fig) and management intensities [30]. All nine parks were regularly watered, but experienced different degrees of floral rotation; some parks periodically brought in new flowering plant species as previously planted species dropped their flowers, while other parks did not rotate plant species at all (A. Stewart, pers. obs.). It is common for parks in Bangkok to have both native and exotic plant species [30].

Data collection

Monthly pollinator observations were conducted over 12 consecutive months (December 2017—November 2018). The order in which parks were visited during each month was random, and sampling was only conducted on sunny days. All sampling at each park was conducted within a single day, and up to three parks were observed in a single day. During each month, the first and last parks were sampled within 16–25 days of each other. During each sampling event, we conducted 15-minute observations at all of the most abundant plant species that were in flower. In order to be efficient with sampling, we only observed plant species with at least 20 flowers or inflorescences; this resulted in approximately 80–90% of the flowering plant community being sampled at each park in each month. We chose this sampling method because we wanted to maximize the number of flowering plant species sampled in order to track floral resource availability across months, and to record as many plant-pollinator interactions as possible, but were limited by manpower. For each plant species observed, a 2 x 2 m plot with abundant flowers of the focal species was selected, and all animals that contacted floral reproductive structures while visiting flowers (i.e., potential pollinators) were recorded. As we did not quantify actual pollen transfer, we cannot confirm that all recorded taxa were true pollinators; it is likely that observed animal taxa differ in pollination effectiveness, and some taxa may not have contributed to pollination at all. We chose to conduct non-destructive pollinator sampling (i.e., relying on visual observation) because collecting individuals may have reduced pollinator abundance and richness at adjacent plots within the park. Unknown pollinators were either photographed or collected with a net, and insects were identified with the help of local field guidebooks and entomologists (see acknowledgments). Plots were not fixed; they varied by month according to which plant species were flowering (mean ± SE plant species sampled per park per month: 13.8 ± 0.7 species; range: 1–33 species). Permission to work with insects was granted by MUSC-IACUC (Faculty of Science, Mahidol University–Institute Animal Care and Use Committee; license number MUSC60-038-388).

We also obtained data on average monthly temperature (hereafter, “temperature”), total monthly precipitation (hereafter, “precipitation”), floral abundance, and floral species richness. Temperature and precipitation data were acquired from the Thai Meteorological Department (www.tmd.go.th). Floral abundance was determined at each plot as the number of flowers per 2 x 2 m plot; these values were then averaged to obtain a mean number of flowers per plot for each park in each month. Floral richness was determined as the number of flowering plant species sampled at each park in each month.

Data analysis

All analyses were conducted in R 3.6.0 [32]. We used linear mixed modelling (LMM; package “lme4”) to examine which predictors influenced pollinator abundance and richness. The response variables examined were total pollinator abundance, total pollinator richness, and the abundance and richness of the most common insect orders (Hymenoptera, Lepidoptera, and Diptera). The fixed factors tested were month, temperature, precipitation, temperature x precipitation, floral abundance, floral richness, and floral abundance x floral richness; park was included as a random factor. QQ plots (package “stats”) were used to check the normality of the residuals. In the model prediction graphs, when two explanatory predictors were found to be significant, one predictor was plotted along the x-axis and the second predictor was represented using various colors; this method also clearly shows when the two predictors have a significant interaction (non-parallel lines) or not (parallel lines).

We also examined pollination networks using the package “bipartite” [33,34]. One overall network was created using data from all parks and all months, and we also created separate networks for each month at each park to examine changes in the pollinator community over time (S2 Fig). We examined network properties at the network level (connectance, weighted connectance, links per species, number of compartments, and Shannon’s diversity), group level (number of species within each trophic level, mean number of links within each trophic level, and niche overlap within each trophic level), and species level (normalized degree averaged across all species, and paired differences index averaged across all species within each trophic level). Descriptions of each network property are provided in S1 Text. We examined whether each network property changed over time using LMM, where month was a fixed factor and park was a random factor. We also calculated species turnover (proportion of species lost or gained between two consecutive months) for both plants and pollinators using the “codyn” package [35].

Results

Pollinator abundance and richness

Over all 12 months of data collection (Fig 1), we observed 8,053 potential pollinators comprising 58 taxa from 4 orders (S1 Table) visiting 136 plant taxa from 48 orders (S2 Table). Hymenoptera were by far the most abundant (95.2%), followed by Diptera (2.5%), Lepidoptera (1.8%), and Hemiptera (0.5%). We observed 23 taxa of Hymenoptera, 23 species of Lepidoptera, 11 taxa of Diptera, and 1 taxon of Hemiptera. The most abundant taxa were Tetragonula stingless bees (60.2%), Apis florea (Fabricius, 1787) (17.6%), A. cerana (Fabricius, 1793) (6.9%), A. dorsata (Fabricius, 1793) (4.9%), Lasioglossum sweat bees (3.7%), and Tephritidae flies (2.1%); all other taxa comprised less than one percent of total abundance (S1 Table).

Fig 1

Temporal variation in weather and pollinator communities in Bangkok, Thailand over 12 months (December 2017—November 2018).

(A) Average monthly temperature (triangles, red line) and total monthly precipitation (circles, blue line). Neither (B) pollinator abundance nor (C) pollinator species richness varied significantly across months; points and error bars represent mean ± SE.

When examining total pollinator abundance, we found that neither month, temperature, precipitation, nor temperature x precipitation had a significant effect, but both floral abundance and floral richness had positive effects, and their interaction was significant as well (Table 1; Fig 2A). For total pollinator richness, both precipitation and floral richness had significant positive effects (Table 1; Fig 2B). Hymenoptera abundance was significantly influenced by floral abundance (positive effect), floral richness (positive effect), and their interaction (Table 1; Fig 2C), while Hymenoptera richness was influenced by precipitation and floral richness (positive effects; Table 1; Fig 2D). Both Lepidoptera abundance and richness were positively influenced by floral richness (Table 1; Fig 2E and 2F). For Diptera abundance and richness, none of the predictors included in the model were significant (Table 1).

Table 1

Results of linear mixed modelling examining the effect of 7 predictors on pollinator abundance and species richness.

	All pollinators		Hymenoptera		Lepidoptera		Diptera
	Abund.Rm2=0.19Rc2=0.40	Rich.Rm2=0.20Rc2=0.52	Abund.Rm2=0.19Rc2=0.04	Rich.Rm2=0.21Rc2=0.42	Abund.Rm2=0.10Rc2=0.34	Rich.Rm2=0.08Rc2=0.36	Abund.n/a	Rich.n/a
Month	χ12=0.09P = 0.76	χ12=0.16P = 0.69	χ12=0.09P = 0.77	χ12=0.43P = 0.51	χ12=2.13P = 0.14	χ12=2.53P = 0.11	χ12=3.43P = 0.06	χ12=3.04P = 0.08
Temperature	χ12=0.22P = 0.64	χ12=0.59P = 0.44	χ12=0.24P = 0.62	χ12=1.33P = 0.25	χ12=0.61P = 0.43	χ12=1.01P = 0.31	χ12=0.03P = 0.87	χ12=0.09P = 0.77
Precipitation	χ12=2.75P = 0.10	χ12=4.23P = 0.04	χ12=2.70P = 0.10	χ12=4.53P = 0.03	χ12=0.24P = 0.62	χ12=0.003P = 0.96	χ12=0.07P = 0.80	χ12=0.04P = 0.84
Temperature x precipitation	χ12=1.29P = 0.26	χ12=0.01P = 0.92	χ12=1.52P = 0.22	χ12=0.20P = 0.66	χ12=1.50P = 0.22	χ12=1.47P = 0.23	χ12=1.29P = 0.26	χ12=0.09P = 0.77
Floral abundance	χ12=6.72P = 0.03	χ12=1.07P = 0.30	χ12=6.77P = 0.03	χ12=1.59P = 0.21	χ12=0.01P = 0.92	χ12=0.23P = 0.63	χ12=1.36P = 0.24	χ12=0.84P = 0.36
Floral richness	χ12=13.7P = 0.001	χ12=8.63P = 0.003	χ12=13.6P = 0.001	χ12=8.90P = 0.003	χ12=6.57P = 0.01	χ12=5.28P = 0.02	χ12=0.21P = 0.65	χ12=0.01P = 0.92
Floral abundancex richness	χ12=4.31P = 0.04	χ12=0.38P = 0.54	χ12=4.39P = 0.04	χ12=0.82P = 0.36	χ12=1.13P = 0.29	χ12=0.47P = 0.49	χ12=0.17P = 0.68	χ12<0.001P = 0.98

Separate analyses were conducted for total pollinator abundance, total pollinator richness, and the abundance and richness of each of the three most common insect orders observed (Hymenoptera, Lepidoptera, and Diptera). Significant predictors are highlighted in yellow with p-values in bold. Marginal () and conditional () R2 values are listed for each final model.

Fig 2

Model predictions of pollinator abundance and richness in Bangkok, Thailand.

(A) Total pollinator abundance was significantly influenced by floral abundance (x-axis), floral richness (denoted by color; red: 6.6 spp., blue: 13.9 spp., green: 21.1 spp.), and their interaction. (B) Total pollinator richness was significantly influenced by floral richness (x-axis) and precipitation (denoted by color; red: 58 mm, blue: 155 mm, green: 251 mm). (C) Hymenopteran abundance was significantly influenced by floral abundance (x-axis), floral richness (denoted by color; red: 6.6 spp., blue: 13.9 spp., green: 21.1 spp.), and their interaction. (D) Hymenopteran richness was significantly influenced by floral richness (x-axis) and precipitation (denoted by color; red: 58 mm, blue: 155 mm, green: 251 mm). (E) Lepidopteran abundance and (F) Lepidopteran richness were significantly influenced by floral richness (x-axis).

Pollination networks

The overall plant-pollinator network (using data from all parks across all months) had a connectance of 0.057, a weighted connectance of 0.092, an average of 2.05 links per species, 3 compartments, and a Shannon diversity index of 4.46. At the group level, pollinators had an average of 5.29 links, while plants had an average of 5.96 links. The average number of shared partners was 0.54 for pollinators and 1.02 for plants. Niche overlap was 0.081 for pollinators and 0.44 for plants. When examining whether network properties changed over time, we found that none of the tested properties varied by month (Table 2; S2 Fig). We did, however, find that both plant and pollinator composition changed across months; species turnover between consecutive months was close to 50% for both plants (mean ± SE: 0.45 ± 0.04) and pollinators (0.46 ± 0.04).

Table 2

Results of linear mixed modelling examining whether plant-pollinator network properties varied across 12 months (December 2017—November 2018).

	Network property	Mean	SE	Chi-square	P
Network level	Connectance	0.356	0.016	χ12=2.29	0.13
	Weighted connectance	0.209	0.008	χ12=2.10	0.15
	Links per species	0.845	0.018	χ12=1.85	0.17
	Number of compartments	1.987	0.106	χ12=1.11	0.29
	Shannon’s diversity	1.796	0.063	χ12=0.64	0.43
Group level	Number of pollinator species	5.228	0.304	χ12=0.40	0.53
	Number of plant species	6.582	0.341	χ12=0.17	0.68
	Links per pollinator species	3.614	0.224	χ12=0.39	0.53
	Links per plant species	1.927	0.095	χ12=2.07	0.15
	Niche overlap among pollinators	0.256	0.026	χ12=0.13	0.72
	Niche overlap among plants	0.433	0.029	χ12=0.52	0.47
Species level	Normalized degree	0.356	0.016	χ12=2.29	0.13
	Paired differences index (pollinators)	0.923	0.007	χ12=2.33	0.13
	Paired differences inex (plants)	0.938	0.009	χ12=0.70	0.40

Network property descriptions are provided in S1 Text. Mean and SE were calculated from 108 plant-pollinator networks (nine parks over 12 months).

Discussion

Neither total pollinator abundance nor richness varied significantly by month. The steady numbers of both individuals and species likely reflect the fact that tropical areas have mild climates, which allows pollinators to be active year-round [8,36]. Yet even tropical environments can have seasonal fluctuations in insect abundance and diversity, because floral (food) availability is often markedly different between the rainy and dry seasons in most natural landscapes [26]. Therefore, the second factor contributing to constant pollinator abundance and richness is likely the constancy of floral resources found in our study parks. Urban parks often favor plant species with showy flowers, and cultivate a mix of native and exotic species that provide abundant floral resources year-round [29,37,38]. Indeed, of the 136 plant taxa observed in this study, 73 species were exotic and 47 were native; the remaining 16 could not be verified (S2 Table). The benefits of reliable floral resources in urban habitats have also been demonstrated in temperate bees [23] and tropical butterflies [39], where the abundance and richness of these taxa were less variable in urban environments than natural habitats.

The factors that significantly influenced total pollinator abundance were floral abundance, floral richness, and their interaction. Therefore, pollinator abundance was greatest when both floral abundance and richness were high, and pollinator abundance remained low (even at high floral abundances) when floral richness was low. It is important to note that our measures of pollinator abundance were obtained from our study plots of flowering plant species, and that many areas of the parks had substantially lower floral abundance; such areas would undoubtedly have fewer pollinators than was observed at our study plots. The importance of floral abundance and richness to pollinators seems to be relatively consistent worldwide and across diverse insect taxa [30,40-45]. Such findings are hardly surprising, given the importance of food resources in supporting pollinator communities [29,30,41-46].

However, our study did not find a significant effect of temperature or precipitation on total pollinator abundance, which is contradictory to several prior studies. For example, Andrade-Silva et al. [47] found that temperature was the most important determinant of euglossine bee abundance in a Brazilian forest, and Silva et al. [48] found that most orders of insects were influenced by temperature in a Brazilian savannah. However, Silva et al. [48] also explained that a likely mechanism behind their results was the increase of leaves and flowers during the transition from the dry to wet season. This justification is also consistent with our results; if insect abundance is primarily driven by food abundance, it seems logical that the year-round supply of flowers in our study parks are able to maintain high pollinator abundance across seasons. We also hypothesize that temperature and water availability in our study area were not variable enough to affect local pollinator abundance. The average monthly temperature during our study period ranged a mere 3 degrees (26.6–29.7°C), and plants within our study parks were never water limited due to regular watering by gardeners.

We also found that total pollinator richness was significantly influenced by floral richness and precipitation. Previous studies in both temperate [49] and tropical [25,50] regions have also found positive correlations between floral diversity and pollinator diversity. It is likely that the higher diversity of flowering plant species attracts a higher diversity of insect pollinators. Less conclusive is the effect of precipitation on pollinator richness. In this study, the rainiest months were September and October, which correspond with the greatest numbers of pollinator species observed. Precipitation likely had little effect on park vegetation and floral resources, given regular watering and park horticulture, but it possibly influenced other aspects of insect life history and/or behavior. Findings from previous studies are mixed. For example, studies of butterflies in Brazil have found the highest species richness to occur during the rainy season [51], during the dry season [25], and during the transition from rainy to dry season [52]. Therefore, it appears that the effects of precipitation on species richness are complex and likely linked to other factors as well, such as plant [51] and predator [52] composition.

When examining the three most common insect orders separately, we found that Hymenoptera results were consistent with the overall results, Lepidoptera abundance and richness were influenced only by floral richness, while Diptera abundance and richness were not explained by any of the tested predictors. Hymenoptera abundance was influenced by the same three factors as total pollinator abundance (floral abundance, floral richness, and their interaction); indeed, Hymenoptera abundance was the driving force behind our results for total pollinator abundance, since Hymenoptera accounted for 95% of all observed pollinators. Lepidoptera were found to forage on fewer plant species than Hymenoptera (S2 Fig), so it seems logical that their foraging is driven by high floral richness, as the presence of numerous floral species would increase the odds of there being at least one species attractive to butterflies. The lack of significant findings for Diptera may be due to small sample sizes, or it is possible that their abundance and richness are influenced by other factors not measured in this study. Previous studies have demonstrated that different species of Diptera [48] can peak during different seasons, resulting in unclear patterns at the level of order.

When examining plant-pollinator networks across 12 consecutive months, we found that network properties remained constant even as plant and pollinator composition changed. Our networks reflect the large majority of plant species found in the parks, although our inability to collect data on the rarest plant species may have influenced some network property estimates. The temporal stability of pollination network properties has been highly investigated in recent years, across a broad range of taxa and landscapes; results vary from reporting temporally stable network properties [10,19-21] to highly dynamic network properties [15,16,18]. We suggest that the stability of network properties depends on both the timescale at which the network is analyzed, as well as the stability of environmental conditions. Networks examined over longer timescales (e.g., over multiple years) are more likely to reveal temporally stable network properties (as found by [10,19-21]), while networks examined over shorter timescales (e.g., within or between seasons) are more likely to find temporally dynamic network properties (as found by [15,16,18]). Moreover, the stability of network properties is likely correlated with the stability of environmental conditions. While our study spanned different seasons, it was conducted in a tropical, urban landscape where both temperature and floral resources remained relatively constant year-round. This reason may explain why we observed no differences in network properties across months, unlike previous year-long or season-long studies conducted in temperate [16] and arctic [15,18] environments.

Conclusions

The findings from this study reveal that tropical urban parks are capable of supporting stable pollinator communities year-round. Even though our study was conducted in Bangkok, a city with little natural vegetation, the city’s parks provide abundant floral resources throughout the year due to landscaping efforts, which favor plant species with showy flowers. However, it is important to note that (1) our networks included only the most abundant plant species, which may over- or underestimate some network properties, and (2) these parks are likely only suitable for pollinator species with generalist foraging and nesting habits, and that are tolerant of human activity. For example, we only observed 58 insect taxa visiting flowers, in contrast to a recent study in Thai mixed fruit orchards that recorded 316 insect pollinator taxa [53]. Moreover, our results suggest that the temporal stability of plant-pollinator network properties is driven by the stability of environmental conditions, including both climate and resource stability. In our study area, the constant floral resources and climate conditions throughout the year appear to create a network in dynamic equilibrium, where plant and pollinator species compositions change, but network properties remain stable year-round. These findings provide insight into how tropical pollinators respond to urban habitats, which will be useful as urban centers continue to grow world-wide.

Map of study parks in Bangkok, Thailand.

(PDF)

Click here for additional data file.

Plant-pollinator networks at each study park in Bangkok, Thailand over 12 months.

(PDF)

Click here for additional data file.

Descriptions of plant-pollinator network properties examined.

(PDF)

Click here for additional data file.

Pollinator species observed in Bangkok, Thailand between December 2017—November 2018.

(PDF)

Click here for additional data file.

Plant species at which pollinator observations were conducted (Bangkok, Thailand; December 2017—November 2018).

(PDF)

Click here for additional data file.

6 Nov 2019

PONE-D-19-19692

Year-round temporal stability of a tropical, urban pollination network

PLOS ONE

Dear Dr. Stewart,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

This work has now been evaluated by two reviewers. Both reviewers find important merits in this study, however they also raised important concerns, and both of them think that the study only partially supports the conclusions and is technically sound. Reviewer I is concerned about the spatial variation of measurements, and raised questions about the type of parks, the study species and the sampling protocols. Reviewer II highlighted several limitations of the methodology, and have raised many concerns about the statistical analyses used. This reviewer also asks for further clarification of the effect of precipitation in parks that are regularly watered, and suggest the addition of references of pollinator abundance or species richness in natural habitats within the same region to evaluate the conservation value of these parks. Based on these evaluations, with which I concur, I would be happy to consider for further evaluation a substantially revised version of the manuscript that considers the suggestions and comments raised by the reviewers. If you decide to do this review, please include a response letter indicating your responses to the reviewer comments and the changes you have made in the manuscript.  If you disagree with a reviewer's point, explain why.

We would appreciate receiving your revised manuscript by Dec 21 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Amparo Lázaro, PhD

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

1. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The study titled “Year-round temporal stability of a tropical, urban pollination network” is interesting and a result well-worthy of publication. The presentation is concise and brings out the important conclusions convincingly. However, there are a few problems that should be addressed for clarity and so that the readers fully understand the implications of the results.

One is a fuller description of the parks. Parks can range from relatively “wild” areas that rely on natural precipitation to highly manicured gardens that are regularly watered. As I read the paper, it became increasing evident that the parks in this study are highly manicured (I think this is correct). This should be conveyed in the introduction and stated again in the methods. If the authors could document how many of the plant species they sampled were native or domesticated, that would give the readers an idea of the types of parks sampled in the study. Watering the plants surely decreased the variation in floral abundance and is partly to explain why the pollination networks change so little through the year.

Second, the title states that the study is about temporal changes over the year, but the comparisons of floral abundance, floral species richness, etc. are conclusions that are based on differences across parks. If this is the case, the study is also about spatial changes. This should be addressed.

Finally, and importantly, is choice of plant species sampled at each park. The text is vague on this and needs to provide more detail. Line 80 states that the plants where observations were made were those that were “the most abundant plant species that were in flower”. How was this determined? How many plants were studied at each park and each date? Why did the study use this approach and not a protocol that sampled the same plant species no matter how many flowers each species had? It is important that the protocol used is justified.

This protocol has important implications for what the measures of floral abundance mean. As q=written it appears that your 2 x 2m plots is extrapolated to the size of the entire park which would result in a gross overestimate of actual floral abundance. Please explain all this more carefully and discuss how this choice influences the interpretation of your results, or it does at all.

Below are more detailed comments

L. 22,23- this relates to one of my larger comments above. This study is about tropical pollination networks that are probably more stable than networks in the temperate zone, but also the study is in highly manage parks. That should be made clear at this point.

L. 29. Add that when “all parks are pooled” there is no variation in pollinator abundance or richness.

L. 30-Add that when parks are compared, there is evidence that floral abundance, etc. influence pollinator richness.

L.80. How did you determine when plants were abundant or rare?

L. 86 swing could be deleted

L 93-95. Please be more explicit and complete about how floral abundance and richness are determined, and extrapolated to total park size, etc. How did you minimize bias?

L109. As stated above, you examined how pollinator communities varied over time (month to month comparisons when all parks are polled), and space (given that the number of flowering plants and floral abundance were compared across parks.

L. 112. The use of the term “trophic level” is confusing given that there is only plants and their pollinators. Should this be compartments, or some other term?

L. 198-199-vaguely written. Rewrite for clarity.

L 225. Change “within” to “of”

L 250. either delete or rewrite the phrase “…that are undeterred by human activity.”

L270. This reference is improperly formatted

Add scale bar to map of parks

Reviewer #2: The presented results give an important and new insight into pollinator communities of tropical cities and I think they should be published. However, data collection had some limitations, limiting also the interpretation of data which should be outlined a little more clearly. First of all, when recording insect visitation to plants, you do not have information on whether the insects were pollinating the plants (even when the insect was observed to touch the reproductive parts of the flower). Therefore, I would caution against using the term pollination network. Instead, I would suggest to rather use the term plant insect/pollinator visitation network.

Furthermore, only the most abundant plants in the parks were observed which means that only a fraction of the entire network was selected which also imposes a strong bias for the assessment of network properties. The use of this method also means that the habitat variables floral abundance and floral richness are actually not representative for the park which could easily be misunderstood by the reader. Floral abundance was based on 2x2 m plots with particularly abundant plants. And floral richness included only those most abundant plants. While these parameters can still be related to observed plant visitors, it should be pointed out more clearly for the interpretation of the results.

Additionally, the analyses with linear mixed modelling should be revised and explained more detailed. Your description is a little unclear, but it sounds like you calculated several independent models for each climatic and habitat variable which, in my opinion, is inadequate. Including month as a random factor in some models, although it was not significant as a fixed factor in another model for the same response variable illustrates part of the problem. All climatic and habitat variables that could potentially affect pollinator abundance or richness should be examined in one model (per response variable). You should also consider using the same explanatory variables (as used for pollinator abundance and richness) for the analyses of the network properties. Using the linear mixed model method (LMM) requires normal distribution of the data. Therefore, you should state whether data was normally distributed. The results of the LMM should also include R2-values. The graphs of model predictions (Fig. 2) show always three curves, for low, medium and high predictions (Fig. 2, A-C), except for one of the graphs (Fig. 2, D). It is not explained what these predictions are based on and why Fig. 2, D has only one curve. Typically, there would be just one curve per response variable and explanatory variable. Please either provide additional information or change the graphs.

Another critical point is the evaluated effect of precipitation. The significance of precipitation as a predictor for any response variable is questionable, because all parks were regularly watered additionally to natural precipitation which the model does not take into account. This influence should be considered in analyses and interpretation.

Finally, you state in your conclusion that the studied urban tropical environments host abundant and diverse bee communities. However, you do not relate pollinator abundance or species richness to pollinator communities of natural habitats within the same region. If possible, this information should be added. Otherwise, the value of these urban habitats remains unclear.

Overall, the manuscript is very nicely written and the clear language makes it easy to follow. I hope that you find my comments helpful for revising your manuscript. I think it will be an interesting and valuable addition to the literature.

Minor comments

Abstract

L32 You mention a change of species composition here in the abstract, but do not include this information in the result section. Please add this in the results.

Introduction

L40 Consider adding another more general reference additional to 1 & 2.

Methods

L71 Maybe number of inhabitants and area of Bangkok could be added.

L73-75 What was the size of the parks?

L79-81 What was the range of number of plant species observed per park or in other words how many different plots did you observe per park? Were plots fixed or newly assigned for each observation?

L87 Within how many days did you observe all nine parks each month? Was the order always the same?

L90 “average” instead of “averagely”

L90-93 As you did not measure temperature yourself, I would rephrase: We also obtained data on… from …

L98-116 Please revise your approach of LMM and add details as suggested above.

Results

L130-133 Please explain what points and bars represent.

L135-142 Please indicate whether the listed significant explanatory variables increase or decrease the response variables.

L145-149 Consider making significant predictors bold for easier reading in black and white.

L151-157 Please explain what low, medium and high predictions are based on or change as suggested above.

L154-156 Graph B does not show precipitation as suggested by the caption. Likewise, graph C does not show floral richness nor the interaction of floral abundance and richness.

Discussion

L195 …most orders of insects…

Conclusion

L245-246 Without a comparison to pollinator communities of natural or other kind of habitats in the same region, it is not really possible to make a statement on whether the recorded pollinator community is very abundant or species rich or how valuable of a habitat the parks are.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Dr. Nicola Seitz

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

11 Dec 2019

Please see the "Response to Reviewers" file.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

24 Jan 2020

PONE-D-19-19692R1

Year-round temporal stability of a tropical, urban pollination network

PLOS ONE

Dear Dr. Stewart,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

This manuscript has been now evaluated by the two previous reviewers. Both of them agree that the revisions of the manuscript have helped to improve considerably the clarity and impact of the paper and that the authors have done a great job with the revision. However, one of the reviewers still have some minor comments and suggestions that I would like the authors to consider. If you disagree with any of the points raised, please explain it in the response letter.

We would appreciate receiving your revised manuscript by Mar 09 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Amparo Lázaro, PhD

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: (No Response)

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript is much clearer and more concise. I have no comments and recommend it for publication.

Reviewer #2: The revised manuscript is much improved and most concerns were addressed and changes performed accordingly. A few unclarities and concerns remain and I would suggest some more minor revisions.

Following up on 1. of my previous comments and the authors reply: Although some studies use the term pollination network, the use of this terminology is problematic and actually incorrect as research has shown that pollination cannot be inferred from flower visitation (see e.g. King et al. 2013 or Popic et al. 2013). Less problematic is the use of plant pollinator networks which I suggest as one alternative to pollination networks and which numerous other studies use (including many you have cited, i.e. Robinson et al. 2018, Alarcón et al. 2008 or CaraDonna et al. 2017). However, the most accurate term is plant insect networks (used e.g. by Vanbergen et al. 2014 or Losapio et al. 2015). For improvements in the fields of pollination ecology and plant insect interactions, I would encourage the use of plant insect networks when visitation and not pollination was assessed like in this study.

Following up on 2. of my previous comments and the authors reply: The sampling methods for determining floral abundance and richness were clarified. However, the resulting bias and limitations for the interpretation of results is still not sufficiently discussed in my opinion. The described selective sampling only from the most abundant flowers creates a bias for network analyses, because plants present in low numbers are consistently missing. Please consider this bias in the discussion. Furthermore, the variables floral abundance and floral richness are not representative for the actual park vegetation as they are based on the most abundant patches of flowers which is a selective measure. Please make this more clear in the discussion to avoid confusion.

L89 It still remains unclear if only one park a day was observed, all in one day or how many parks per day within what time span of a month (e.g. three parks a day, and all parks within first week of the month).

L96 visitation interactions instead of pollination interaction

L112-113 If only those flowers were observed that had a minimum of 20 flowers/inflorescences this calculation does not really reflect the floral abundance of the park.

L164-167 Suggestion for consistency/easy reading: either individual brackets with “(positive effect)” always after the first and second predictor or “(positive effects)” only after the second predictor.

L180-189/Fig. 2: I find the interpretation of the graphs still a little difficult. Can you provide more information on the visualization of the model predictions in the methods section?

L272-278 As mentioned above, you should point out that the visitation networks analyzed completely omitted less abundant plants which is potentially affecting network properties.

L289-290 Maybe this could be rephrased slightly, because considering the much lower species diversity compared to the below mentioned fruit orchards, the pollinator community in the parks does not appear to have a very high level of diversity.

References:

King, C., Ballantyne, G. and Willmer, P.G. (2013), Why flower visitation is a poor proxy for pollination: measuring single‐visit pollen deposition, with implications for pollination networks and conservation. Methods Ecol Evol, 4: 811-818. doi:10.1111/2041-210X.12074

Popic, T.J., Wardle, G.M. and Davila, Y.C. (2013), Flower‐visitor networks only partially predict the function of pollen transport by bees. Austral Ecology, 38: 76-86. doi:10.1111/j.1442-9993.2012.02377.x

Robinson, Samuel VJ, Gianalberto Losapio, and Gregory HR Henry. "Flower-power: Flower diversity is a stronger predictor of network structure than insect diversity in an Arctic plant–pollinator network." Ecological complexity 36 (2018): 1-6.

Alarcón, Ruben, Nickolas M. Waser, and Jeff Ollerton. "Year‐to‐year variation in the topology of a plant–pollinator interaction network." Oikos 117.12 (2008): 1796-1807.

CaraDonna, P. J., Petry, W. K., Brennan, R. M., Cunningham, J. L., Bronstein, J. L., Waser, N. M., & Sanders, N. J. (2017). Interaction rewiring and the rapid turnover of plant–pollinator networks. Ecology letters, 20(3), 385-394.

Vanbergen, A.J., Woodcock, B.A., Gray, A., Grant, F., Telford, A., Lambdon, P., Chapman, D.S., Pywell, R.F., Heard, M.S. and Cavers, S. (2014), Grazing alters insect visitation networks and plant mating systems. Funct Ecol, 28: 178-189. doi:10.1111/1365-2435.12191

Losapio, G., Jordán, F., Caccianiga, M., & Gobbi, M. (2015). Structure-dynamic relationship of plant–insect networks along a primary succession gradient on a glacier foreland. Ecological modelling, 314, 73-79.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Dr. Nicola Seitz

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

28 Jan 2020

Please see the uploaded "Response to Reviewers" file.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

3 Mar 2020

Year-round temporal stability of a tropical, urban plant-pollinator network

PONE-D-19-19692R2

Dear Dr. Stewart,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Amparo Lázaro, PhD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

25 Mar 2020

PONE-D-19-19692R2

Year-round temporal stability of a tropical, urban plant-pollinator network

Dear Dr. Stewart:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Amparo Lázaro

Academic Editor

PLOS ONE