https://doi.org/10.5061/dryad.1c59zw427

Datasets
1. "ML_interactions_dna_barcoding_samples.csv"
2. "ML_networks_per_landscape_habitat.csv"
3. "ML_disperser_traits.csv"
4. "ML_bird_migration.csv"
5. "ML_plant_traits.csv"

Description of the data and file structure

1. "ML_interactions_dna_barcoding_samples.csv"
Data on the samples of bird droppings and mammal scats for which we successfully identified the bird species that dispersed the seeds through DNA-barcoding analysis. The dataset includes 3093 rows for 3063 samples because some samples (droppings or scats) included more than one seed species. That is, 3063 samples included 3093 interaction events between a frugivore-plant species pair, and contained a total of 14,683 seeds.

Variables:

1. sample_code: sample level codes.
2. landscape: study landscape (seven levels).
3. date: sampling date.
4. seed_sp: seed species identified visually (see Table S4). In 11 out of the 3063 samples DNA barcoding also contributed to seed species identification.
5. nseeds: number of seeds per plant species in the sample.
6. habitat: habitat of seed deposition, where the samples were collected (two levels: ‘forest’ and ‘matrix’).
7. mhabitat: microhabitat type of seed deposition, where the samples were collected (four levels: ‘tree’, ‘shrub’, ‘pylon’ and ‘open’).
8. perch: perch of seed deposition (three levels: ‘canopy’, ‘pylon’ and ‘open’ for seed deposition in open spaces).
9. perch_type: perch type of seed deposition (two levels: ‘natural’ and ‘pylon’ for anthropogenic perches).
10. sampling_type: sampling type for sample collection (three levels: ‘tray’ for sampling in seed traps, ‘transect’ for sampling within belt transects, and ‘observed’ for samples found visually outside seed traps and transects).
11. disperser_group: group of seed disperser (two levels: ‘bird’ and ‘mammal’).
12. disperser_sp: bird and mammal species identified through DNA barcoding, following taxonomy from the ‘Birds of the World’ (www.birdsoftheworld.org) and the ‘Integrated Taxonomic Information System’ for mammals (www.itis.gov).
13. disperser_sp_original: bird and mammal species identified in BOLD or BLAST.
14. identification: whether the seed disperser that ejected the sample (dropping or scat) was successfully identified through DNA barcoding (‘yes’: 3090 out of the 3093 samples) or was inferred (‘inferred’: 3 mammal scats) from DNA barcoding results of similar samples (i.e. mammal scats) from the same landscape.
15. sequence: nucleotide sequence obtained through DNA barcoding analysis to identify the disperser species.
16. similarity: percentage of similarity with best matching sequence in BOLD or BLAST.
17. genbank_anumber: accession number of best matching sequence deposited in GenBank; ‘Private’ denotes private sequences from BOLD, whereas codes starting with ‘BIN_Id_in_BOLD’ denote ‘Barcode Index Numbers’ from BOLD.
18. similarity_best_match_genbank: percentage of similarity with best matching sequence in GenBank, provided when best matched sequences in BOLD were ‘Private’.
19. top_genbank_anumber_if_private: accession number of best matching sequence deposited in GenBank when best matching sequence in BOLD were ‘Private’.
20. plant_sequence: nucleotide sequence obtained through DNA barcoding analysis to identify the seed species.

________________________

2. "ML_networks_per_landscape_habitat.csv"
Data on seed-dispersal networks between frugivore species and the seed species they dispersed in the forest and matrix of the study landscapes. The dataset includes 401 rows, that is, interactions between species pairs in each habitat type and study landscape (range = 44–83 pairwise interactions per landscape). The interaction weight of pairwise interactions was quantified as the seed-rain density (dispersed seeds per m²) of plant species i dispersed by frugivore species j in the forest and matrix of each study landscape. These data result from combining seed-rain data measured in seed traps and transects and the contribution of frugivore species to the seed rain of each plant species (extracted from ‘ML_interactions_dna_barcoding_samples.csv’) as explained in the Methods section and the SI Appendix (Appendix S1) of the Supporting Information.

Variables:

1. landscape: study landscape (seven levels).
2. habitat: habitat of seed deposition, where the samples were collected (two levels: ‘forest’ and ‘matrix’).
3. disperser_group: group of seed disperser (two levels: ‘bird’ and ‘mammal’).
4. disperser_sp: bird and mammal frugivorous species that dispersed the seeds.
5. seed_sp: dispersed seed species.
6. ism2: interaction-level seed rain quantifying the weight of pairwise interactions as the seed-rain density (sr_ij, expressed as seeds per m²) of plant species i dispersed by frugivore species j in the forest and matrix of each study landscape.

________________________

3. "ML_disperser_traits.csv"
Species-level data on body mass and hand-wing index (HWI) for the 43 frugivore species (43 rows) identified as seed dispersers after DNA-barcoding analyses of bird droppings and mammal scats containing seeds. Taxonomy from the ‘Birds of the World’ (www.birdsoftheworld.org) and the ‘Integrated Taxonomic Information System’ for mammals (www.itis.gov). Trait data was extracted from AVONET database (https://onlinelibrary.wiley.com/doi/full/10.1111/ele.13898) for birds and from EltonTraits 1.0 (https://esajournals.onlinelibrary.wiley.com/doi/abs/10.1890/13-1917.1) for mammals.
 
Variables:

1. disperser_sp: animal species.
2. disperser_group: group of seed disperser (two levels: ‘bird’ and ‘mammal’).
3. order: taxonomic Order.
4. family: taxonomic Family.
5. body_mass: mean body mass of the species in grams.
6. HWI: hand-wing index of bird species expressed as percentages (HWI = 100 × Kipp’s distance / wing length). For mammals, HWI = NA.
7. data_source: database from which the traits were extracted.

________________________

4. "ML_bird_migration.csv"
Migration data for the 40 bird species that acted as seed dispersers (SI Appendix, Table S3) at the study landscapes (101 rows) because the migration traits vary geographically within species. The data were extracted from González-Varo et al. (2021; https://doi.org/10.1038/s41586-021-03665-2), which collated data from multiple data sources (see this article for details; data including original sources available at https://doi.org/10.5061/dryad.15dv41nx3).
 
Variables:

1. landscape: study landscape (seven levels).
2. disperser_sp: bird species.
3. prop_migrants: variable characterizing the proportion of migrants (P_migrants) of the frugivorous bird species at the study landscapes: 0, non-migrant population; 0.1, only a minor fraction migrates; 0.25, a larger fraction migrates but non-migrants prevail; 0.5, roughly half of the population migrates; 0.75, migrants prevail; 0.9, only a minor fraction does not migrate; 1: the whole population migrates.
4. migrant_type: type of migrant of the bird species in the study landscape (three levels: ‘non_migrant’ for resident birds; ‘short_distance’ for Palearctic migrants; ‘long_distance’ for Afro-Palearctic migrants).

________________________

5. "ML_plant_traits.csv"
Trait data for the 48 plant species (SI Appendix, Tables S4 and S5) at the different study landscapes (99 rows) because phenological data vary geographically within species. Mean seed weight and plant height were obtained at the plant species level, whereas phenological data was obtained for the plant species at the bioclimate of the study landscapes. The data sources for seed mass and plant height are detailed in SI Appendix, Table S5. Phenological data was obtained from González-Varo et al. (2021; https://doi.org/10.1038/s41586-021-03665-2).
 
Variables:

1. landscape: study landscape (seven levels).
2. seed_sp: seed species (see Table S4).
3. seed_weight_mg: mean seed weight in milligrams.
4. plant_height_m: mean plant height in meters.
5. exotic_cultivated_planted: Bernoulli-distributed variable to classify the seed species according to the origin of their adult plants in each landscape (1: exotic or planted; 0: wild and native).
6. fruit_L: The average start date (d_start) of the fruiting (seed-dispersal) period expressed on a monthly scale.
7. fruit_R: The average end date (d_end) of the fruiting (seed-dispersal) period expressed on a monthly scale.

Sharing/Access information

Trait data for frugivores and plants were obtained from the following sources:

• https://datadryad.org/stash/dataset/doi:10.5061/dryad.15dv41nx3
• https://onlinelibrary.wiley.com/doi/full/10.1111/ele.13898
• https://esajournals.onlinelibrary.wiley.com/doi/abs/10.1890/13-1917.1
• https://onlinelibrary.wiley.com/doi/10.1111/gcb.14904 <br>

Code/Software

R project and code to reproduce the data analyses is available at zenodo (R_Project_Mobile_Links).
The code consists of the following six files:

code_1_sample_coverage_completeness.R
code_2_alpha_beta_diversity.R
code_3_H2_complementarity.R
code_4_disperser_groups.R
code_5_cwms_trait_analysis.R
code_6_beta_cwms_correlations.R

All analyses were conducted in the R using the software versions listed below.

R version 4.1.2 (2021-11-01)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Big Sur 11.6.4
Matrix products: default
LAPACK: /Library/Frameworks/R.framework/Versions/4.1/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats graphics grDevices utils datasets methods base

other attached packages:
[1] gjam_2.6.2 bipartite_2.17 sna_2.6
[4] network_1.17.1 statnet.common_4.5.0 vegan_2.6-2
[7] lattice_0.20-45 permute_0.9-7 DHARMa_0.4.5
[10] MuMIn_1.46.0 performance_0.9.0 glmmTMB_1.1.3
[13] emmeans_1.7.3 cowplot_1.1.1 reshape2_1.4.4
[16] tidyr_1.2.0 betapart_1.5.6 ggplot2_3.3.5
[19] iNEXT_2.0.20

loaded via a namespace (and not attached):
[1] TH.data_1.1-1 minqa_1.2.4 colorspace_2.0-3
[4] ellipsis_0.3.2 estimability_1.3 rstudioapi_0.13
[7] farver_2.1.0 fansi_1.0.3 mvtnorm_1.1-3
[10] codetools_0.2-18 splines_4.1.2 doParallel_1.0.17
[13] itertools_0.1-3 spam_2.8-0 nloptr_2.0.0
[16] cluster_2.1.3 shiny_1.7.2 compiler_4.1.2
[19] Matrix_1.5-1 fastmap_1.1.0 cli_3.6.1
[22] later_1.3.0 htmltools_0.5.2 tools_4.1.2
[25] dotCall64_1.0-1 igraph_1.3.1 coda_0.19-4
[28] gtable_0.3.0 glue_1.6.2 RANN_2.6.1
[31] dplyr_1.0.7 maps_3.4.0 fastmatch_1.1-3
[34] Rcpp_1.0.10 carData_3.0-5 vctrs_0.6.1
[37] ape_5.6-2 nlme_3.1-157 iterators_1.0.14
[40] insight_0.18.6 stringr_1.4.0 lme4_1.1-29
[43] mime_0.12 lifecycle_1.0.3 gap.datasets_0.0.5
[46] MASS_7.3-57 zoo_1.8-10 scales_1.2.0
[49] rcdd_1.5 doSNOW_1.0.20 promises_1.2.0.1
[52] parallel_4.1.2 sandwich_3.0-2 TMB_1.8.1
[55] fields_13.3 see_0.7.0 gridExtra_2.3
[58] stringi_1.7.6 bayestestR_0.12.1 gap_1.2.3-6
[61] foreach_1.5.2 boot_1.3-28 geometry_0.4.6
[64] qgam_1.3.4 rlang_1.1.0 pkgconfig_2.0.3
[67] purrr_0.3.4 patchwork_1.1.1 labeling_0.4.2
[70] tidyselect_1.1.2 plyr_1.8.7 magrittr_2.0.3
[73] R6_2.5.1 snow_0.4-4 generics_0.1.2
[76] multcomp_1.4-19 picante_1.8.2 pillar_1.7.0
[79] withr_2.5.0 mgcv_1.8-40 survival_3.3-1
[82] datawizard_0.4.0 abind_1.4-5 tibble_3.1.6
[85] crayon_1.5.1 car_3.0-12 utf8_1.2.2
[88] viridis_0.6.2 grid_4.1.2 digest_0.6.29
[91] xtable_1.8-4 httpuv_1.6.5 numDeriv_2016.8-1.1
[94] stats4_4.1.2 munsell_0.5.0 viridisLite_0.4.0
[97] magic_1.6-0