Timing and social dynamics of divorce in wild great tits: A phenomenological approach
Data files
Jul 08, 2025 version files 1.89 GB
-
Analysis_code.R
92.66 KB
-
birds_flocking_events_1.csv
90.66 MB
-
birds_flocking_events_3.csv
52.83 MB
-
event_information_1.csv
7.37 MB
-
event_information_2.csv
7.45 MB
-
event_information_3.csv
6.14 MB
-
birds_flocking_events_2.csv
75.78 MB
-
individual_records_1.csv
101.24 KB
-
individual_records_2.csv
86.30 KB
-
individual_records_3.csv
77.54 KB
-
pair_statuses_allyears.csv
524.26 KB
-
rand_assoc.csv
22.50 MB
-
README.md
13.56 KB
-
records_events_list_1.csv
610.94 MB
-
records_events_list_2.csv
583.20 MB
-
records_events_list_3.csv
436.66 MB
Abstract
Social behaviour is a key part of life for many species. In monogamously breeding species, social associations between breeding partners are particularly important. Selection of a breeding partner often begins well before reproduction, and this process can affect subsequent reproductive success. Thus, the non-breeding season can shape behaviour during the breeding season. However, it is currently unknown how breeding season outcomes can impact associate choice during the non-breeding season, as studying this requires high volumes of cross-context individual social data. This study used three years of wild great tit social data from Wytham Woods, Oxford, UK, to examine social associations between pairs classified with respect to prior and future breeding status (divorcing, faithful, new and juvenile). We found distinct patterns of social association in ’divorcing’ pairs from early winter, suggesting that divorce is an ongoing behavioural process. Newly forming pairs initially associated similarly to divorcing pairs, but became similar to faithful pairs over time. On a finer spatiotemporal scale, the behaviour of faithful and divorcing birds diverged over the winter. These results provide the first evidence of a behavioural signature of divorce during the non-breeding season in great tits, while suggesting that different behaviours may drive behavioural divorce at different times.
Abraham, A. D.; Sheldon, B. C.; Firth, J. A. Manuscript DOI: 10.1098/rspb.2024.3065
Dataset DOI: 10.5061/dryad.vq83bk453
Description of the data and file structure
Three years of wild great tit social data were used to analyse how winter (non-breeding) behaviour differs between breeding pairs that do and do not remain together for consecutive springs. This analysis was conducted at two different spatiotemporal scales: the number of shared flocking events at supplementary feeders, and how often pairs came to the feeder immediately before or after each other within flocking events that they shared.
All data analysis was done in R Version 4.2.2.
Files and variables
File: Analysis_code.R
Description: The code used to analyse the data and produce results and plots included in the manuscript.
File: event_information_1.csv File: event_information_2.csv File: event_information_3.csv
Description:
These CSV files contain each flocking event calculated from data used in the analysis, including information about flock size, location, and time. Flocking events were calculated using the asnipe package.
CSV 1 contains records from winter 2011/12, CSV 2 from winter 2012/13, and CSV 3 from winter 2013/14. Columns are the same across each file.
Column summary:
- flockevent
Flocking event number
- location
Feeder id where the flocking event occurred
- start.time
Time when the flocking event was estimated to begin
- stop.time
Time when the flocking event was estimated to end
- flocksize
Number of tagged birds observed in the flocking event
- day
Day of the year that the flocking event was recorded (e.g., February 2 = 33)
- year
Data year in which the flocking event was recorded
Data years are calculated as follows
Winter 2011/12 = 2012
Winter 2012/13 = 2013
Winter 2013/14 = 2014
File: birds_flocking_events_1.csv File: birds_flocking_events_2.csv File: birds_flocking_events_3.csv
Description:
These CSV files provide a record of every flocking event a bird was a part of, along with the location, date, and time of that event.
CSV 1 contains records from winter 2011/12, CSV 2 from winter 2012/13, and CSV 3 from winter 2013/14. Columns are the same across each file.
Column summary:
- ring
Bird BTO ring number
- flockevent
Flocking event number (unique within but not between data years)
- location
Feeder the flocking event occurred at
- start_time
Time the flocking event began
- stop_time
Time the flocking event ended
- nwkend
Weekend of the year that the flocking event occurred in
- arrive
Was the individual’s flocking event before this one at a different feeder? (Binary TRUE/FALSE). NAs in this column occur at the first record at a feeder on a given day, as we are unable to say whether the bird was at a different feeder before that record.
- leave
Was the individual’s flocking event after this one at a different feeder? (Binary TRUE/FALSE). NAs in this column occur at the last record at a feeder on a given day, as we are unable to say whether the bird was at a different feeder after that record.
- day
Julian day that the flocking event occurred
- year
Data year in which the flocking event was recorded
Data years are calculated as follows
Winter 2011/12 = 2012
Winter 2012/13 = 2013
Winter 2013/14 = 2014
- exday
‘Experimental day’ - day of data being recorded during that data year. As recording is only on weekends and started at slightly different times across years, the relationship between exday and day will differ between data years.
File: individual_records_1.csv File: individual_records_2.csv File:individual_records_3.csv
Description:
Each of these CSV files contains demographic information for each individual observed across the years of data collection.
CSV 1 contains records from winter 2011/12, CSV 2 from winter 2012/13, and CSV 3 from winter 2013/14. Columns are the same across each file.
Column summary:
- id
BTO ring number of the individual
- species
Species codes are as follows:
‘bluti’ = Blue tit
‘greti’ = Great tit
‘coati’ = Coal tit
‘marti’ = Marsh tit
‘nutha’ = Nuthatch
Only great tit records are used in this analysis
- sex
The individual’s sex (M, F, or U (unknown))
- age
Estimated age of the individual from ringing records
- immigrant
‘FALSE’ if the bird was ringed as a chick in Wytham Woods, ‘TRUE’ otherwise
File: records_events_list_1.csv File: records_events_list_2.csv File: records_events_list_3.csv
Description:
These CSV files contain each record of a tagged bird at a feeder, along with additional information, such as the flocking event of that record and preceding/following birds, to help with analysis and cross-reference.
CSV 1 contains records from winter 2011/12, CSV 2 from winter 2012/13, and CSV 3 from winter 2013/14. Columns are the same across each file.
Column summary:
- id
ID number for each individual bird
- location
Feeder ID where the visit was recorded
- time
Time at which the visit was recorded
- day
Day of the year that the visit was recorded (e.g., February 2 = 33)
- year
Data year in which the visit was recorded
Data years are calculated as follows
Winter 2011/12 = 2012
Winter 2012/13 = 2013
Winter 2013/14 = 2014
- flockevent
Flocking event that this record is classified in, to allow cross-referencing with other data.
- lastbird
ID of the bird observed at the relevant feeder in the record before. NAs in this column occur at the first record at a feeder on a given day, as we are unable to say which bird was at the feeder before that record.
- nextbird
ID of the bird observed at the relevant feeder in the record after. NAs in this column occur at the last record at a feeder on a given day, as we are unable to say whether the bird was at a different feeder after that record.
- arrive
Is this the first consecutive record of this bird at this feeder? I.e has this bird just ‘arrived’ (binary TRUE/FALSE). NAs in this column occur at the first record at a feeder on a given day, as we are unable to say whether the bird was at a different feeder before that record.
- leave
Is this the last consecutive record of this bird at this feeder? I.e after this record does the bird ‘leave’ (binary TRUE/FALSE). NAs in this column occur at the last record at a feeder on a given day, as we are unable to say whether the bird was at a different feeder after that record.
- exday
‘Experimental day’ - day of data being recorded during that data year. As recording is only on weekends and started at slightly different times across years, the relationship between exday and day will differ between data years.
File: pair_statuses_allyears.csv
Description:
This data frame provides the foundation for the pairs which are used for analysis, as well as allowing a more detailed demonstration of how pair statuses are estimated.
Column summary:
- male.id
ID of the male bird within the pair
- female.id
ID of the female bird within the pair
- pair.id
Unique identifier for pairs (consistent across years) created by pasting together the male id and female id
- year
Data year for which pair status is being calculated. For the breeding season, data year is the spring in which the pair was observed (i.e spring 2012 = 2012)
- pairobs.year
Whether the pair bred together during the data year
- maleobs.year
Whether the male was observed breeding during that year
- femaleobs.year
Whether the female was observed breeding during that year
- pairobs.prev
Whether the pair bred together during the previous year
- maleobs.prev
Whether the male was observed breeding during the previous year
- femaleobs.prev
Whether the female was observed breeding during the previous year
- male.juv
Is the male a juvenile (born in the previous year)?
- female.juv
Is the female a juvenile (born in the previous year)?
- male.later
Is the male observed in any year after this one (do we know he is not dead)?
- female.later
Is the female observed in any year after this one (do we know she is not dead)?
- status
Defined pair status (see manuscript supplementary material for a description of status options)
- simplestatus
Statuses condensed into fewer, more usable categories (categories used in final analysis)
- maledata.prev
Is there usable feeder record data for the male in the winter before the relevant spring? NAs in this column are used for birds in pairs in the springs of 2011 or 2015 which did not have a feeder experiment run during the previous winter.
- femaledata.prev
Is there usable feeder record data for the female in the winter before the relevant spring? NAs in this column are used for birds in pairs in the springs of 2011 or 2015 which did not have a feeder experiment run during the previous winter.
- maledata.next
Is there usable feeder record data for the male in the winter following the relevant spring? NAs in this column are used for birds in pairs in the springs of 2014 or 2015 which did not have a feeder experiment run during the following winter.
- femaledata.next
Is there usable feeder record data for the female in the winter following the relevant spring? NAs in this column are used for birds in pairs in the springs of 2014 or 2015 which did not have a feeder experiment run during the following winter.
- pairdata.prev
Is there usable feeder record data for both birds in the winter before the relevant spring? NAs in this column are used for pairs in the springs of 2011 or 2015 which did not have a feeder experiment run during the previous winter.
- pairdata.next
Is there usable feeder record data for both birds in the winter following the relevant spring? NAs in this column are used for pairs in the springs of 2014 or 2015 which did not have a feeder experiment run during the following winter.
File: rand_assoc.csv
Description:
This dataframe provides a list of, for each focal individual used in the analysis, every individual they were observed sharing a flocking event with.
Column summary:
- id
Focal bird BTO ring number
- pairnum
Unique identifier for the pair which the focal individual is a part of
- year
Data year for the pair and relevant winter.
Data years are calculated as follows
Winter 2011/12 = 2012
Winter 2012/13 = 2013
Winter 2013/14 = 2014
- original_pairid
The BTO ring number of the bird in the focal bird’s original analysis pair.
- sex
The focal bird’s sex (M, F, or U)
- pairtype
‘original’ = the pair recorded in this row is an original pair, used in the other analyses
‘random’ = the pair recorded in this row is the focal bird and a non-pair associate
- rand_pairid
BTO ring number of the non-pair associate. NAs are used when the pair is original, so there is no non-pair associate.
- day
Julian day that the association was observed
- maxpairs
The number of non-pair associates the focal bird had on that day
- final_pairid
The ‘original_pairid’ and ‘rand_pairid’ columns merged, recording the relevant partner of the focal bird for that row with no NAs.
- exday
‘Experimental day’ - day of data being recorded during that data year. As recording is only on weekends and started at slightly different times across years, the relationship between exday and day will differ between data years.
- sharedevents
Number of flocking events the focal bird and partner (final_pairid) were observed in together on that day
- focal.n.events
Number of flocking events the focal bird was observed in on that day
- partner.n.events
Number of flocking events the partner was observed in on that day
- simplestatus
Pair status of the ‘original’ pair
- sri
Winter association score as calculated with sharedevents, focal.n.events, and partner.n.events
- failures
partner.n.events + focal.n.events + sharedevents, to be used in modelling
Code/software
Commented code is available in Analysis_code.R. All analysis was completed in R version 4.2.2, using the following additional packages:
tidyverse (Version 2.0.0: https://www.tidyverse.org)
glmmTMB (Version 1.1.9: https://github.com/glmmTMB/glmmTMB)
data.table (Version 1.14.8: https://r-datatable.com)
DHARMa (Version 0.4.6: http://florianhartig.github.io/DHARMa/)
ggeffects (Version 1.5.1: https://strengejacke.github.io/ggeffects)
AICcmodavg (Version 2.3-1)
igraph (Version 1.3.5: https://igraph.org)
sna (Version 2.7-1: https://statnet.org)
asnipe (Version 1.1.16)
vegan (Version 2.6-4: https://github.com/vegandevs/vegan)
glmnet (Version 4.1-8: https://glmnet.stanford.edu)
sf (Version 1.0-9: https://r-spatial.github.io/sf/)
geosphere (Version 1.5-18)
terra (Version 1.7-3: https://rspatial.org/terra/)
janitor (Version 2.2.0: https://github.com/sfirke/janitor)
ape (Version 5.7-1: http://ape-package.ird.fr/)
semTools (Version 0.5-6: https://github.com/simsem/semTools/wiki)
ggpubr (Version 0.6.0: https://rpkgs.datanovia.com/ggpubr/)