Data from: Identifying timescales of change in vulture social networks
Data files
Dec 22, 2025 version files 473.18 MB
-
data_seasons_simplified_shifted_1.csv
22.72 MB
-
data_seasons_simplified_shifted_2.csv
25.73 MB
-
data_seasons_simplified_shifted_3.csv
37.67 MB
-
data_seasons_simplified_shifted_4.csv
34.07 MB
-
data_seasons_simplified_shifted_5.csv
57.99 MB
-
data_seasons_simplified_shifted_6.csv
71.88 MB
-
data_seasons_simplified_shifted_7.csv
43.61 MB
-
data_seasons_simplified_shifted_8.csv
74.04 MB
-
data_seasons_simplified_shifted_9.csv
88.41 MB
-
graphs_feeding_seasons.RDS
484.39 KB
-
graphs_feeding.RDS
4.57 MB
-
graphs_flight_seasons.RDS
511.94 KB
-
graphs_flight.RDS
11.33 MB
-
README.md
5.09 KB
-
roostPolygons_shifted.kml
162.89 KB
Abstract
Animal social interaction patterns change over time, but the continuous nature of social interactions makes selecting a timescale for temporal analysis challenging. We applied both a heuristic approach and a multilayer reducibility analysis approach to study timescales of change in social networks of free-ranging griffon vultures. We analyzed social networks in two behavioral situations: in-flight interactions, such as during foraging movements, which we expected to fluctuate seasonally but to exhibit a relatively constant pattern of change over the course of a season; and diurnal ground interactions, such as interactions while feeding, which we expected to show a pulsed temporal pattern that followed the pattern of carcass availability on the landscape. The heuristic method confirmed the suitability of a 3-10 day aggregation window for studying temporal change in vulture social networks. It also highlighted how different timescales of aggregation offer different insights about longer-term patterns of change. Multilayer reducibility analysis confirmed that substantial change was happening at every aggregation timescale we tested, with no redundancy in network layers; that is, social interactions in this population were not oversampled. However, it revealed more similarity between non-adjacent layers in the flight networks as compared to the feeding networks, further supporting the influence of carcasses as drivers of social network structure. Multilayer reducibility analysis over a multi-season timescale did not reveal seasonal similarities in network structure extensive enough to override the substantial demographic and tag coverage turnover between seasons. While multilayer analysis of temporal dynamics in social networks may prove useful for the study of change in links among a fixed subset of individuals, we highlight its limitations as a tool for studying long-term social network structural change, especially in free-living animal populations.
Dataset DOI: 10.5061/dryad.mpg4f4rcf
Description of the data and file structure
Data associated with the paper "Identifying timescales of change in vulture social networks", by Kaija Gahm, Elvira D'Bastiani, Nili Anglister, Gideon Vaadia, Marta Acácio, Orr Spiegel, and Noa Pinter-Wollman.
Collected from GPS-tagged griffon vultures (Gyps fulvus).
For details on how the GPS data were cleaned and processed, see the methods and supplementary material for Gahm et al. (link to preprint)
For details on how the roost locations were defined, see the methods for Acacio et al. 2024
Files and variables
File: roostPolygons_shifted.kml
Description: Polygons used to define roost areas. Used in this analysis to exclude social interactions that occurred during the day inside these areas, to distinguish probable feeding interactions from in-roost diurnal ground interactions. These polygons have been shifted in the X and Y directions by a randomly-selected amount to protect the location of this sensitive species.
File: data_seasons_simplified_shifted_1.csv
Description: Season 1 (Pre-breeding 2020) movement data used to derive proximity-based social interactions.
Variables (these descriptions apply to subsequent seasons as well)
- ground_speed: speed at which the vulture was moving, in m/s. Used to threshold for flying/stationary points
- location_lat: latitude
- location_long: longitude
- timestamp: timestamp of the GPS fix
- dateOnly: date of the GPS fix
- Nili_id: ID (name) of the vulture
- seasonUnique: Full name of the season (e.g. 2020 fall, 2023 summer)
- geometry: SF geometry columns. To be used when converting data back into sf objects.
File: data_seasons_simplified_shifted_2.csv
Description: Season 2 (Breeding 2020-2021) movement data used to derive proximity-based social interactions.
File: data_seasons_simplified_shifted_3.csv
Description: Season 3 (Post-breeding 2021) movement data used to derive proximity-based social interactions.
File: data_seasons_simplified_shifted_4.csv
Description: Season 4 (Pre-breeding 2021) movement data used to derive proximity-based social interactions.
File: data_seasons_simplified_shifted_5.csv
Description: Season 5 (Breeding 2021-2022) movement data used to derive proximity-based social interactions.
File: data_seasons_simplified_shifted_6.csv
Description: Season 6 (Post-breeding 2022) movement data used to derive proximity-based social interactions.
File: data_seasons_simplified_shifted_7.csv
Description: Season 7 (Pre-breeding 2022) movement data used to derive proximity-based social interactions.
File: data_seasons_simplified_shifted_8.csv
Description: Season 8 (Breeding 2022-2023) movement data used to derive proximity-based social interactions.
File: data_seasons_simplified_shifted_9.csv
Description: Season 9 (Post-breeding 2023) movement data used to derive proximity-based social interactions.
File: graphs_feeding_seasons.RDS
RDS file containing season-scale social networks for the feeding situation. Can be loaded into R with readRDS()
File: graphs_flight_seasons.RDS
RDS file containing season-scale social networks for the flight situation. Can be loaded into R with readRDS()
File: graphs_feeding.RDS
RDS file containing sub-season scale social networks (summer/post-breeding 2023) for the feeding situation. Can be loaded into R with readRDS()
File: graphs_flight.RDS
RDS file containing sub-season scale social networks (summer/post-breeding 2023) for the flight situation. Can be loaded into R with readRDS()
Code/software
Code associated with this manuscript can be found at github.com/kaijagahm/vultureTemporal.
The data used in the analysis is prepared using the targets pipeline, which is outlined in _targets.R. The data files provided here are not the original raw data, but they represent various steps in the pipeline. To reproduce the analysis, you could edit the targets pipeline or follow it in a new script.
roostPolygons_shiftedwill take the place ofroostPolygonsin the targets pipeline.- the data files
data_seasons_simplified_shifted_*need to be read into R (e.g. using read_csv()) and then bound together into a list, in order. Then that list will take the place ofalldata. - Alternatively, you can read in the social networks directly from the four RDS files provided, skipping the preliminary analysis steps and proceeding straight to the multilayer reducibility analyses.
It is possible that some analyses may not run seamlessly with the shifted data. Access to the original, un-shifted data will be available upon request.