Data from: An urban bird population exhibits more fear in the non-breeding season
Data files
Apr 30, 2026 version files 67.55 KB
-
README.md
4.73 KB
-
Seasonal_Fear_Code_Dryad.Rmd
6.26 KB
-
Seasonal_Fear_Data_Jan29.csv
56.56 KB
Abstract
Fear response in birds is affected by environmental factors, many of which are related to seasonality, such as temperature or resource availability. The evolutionarily novel conditions of the urban environment may alter avian seasonal behaviours. We ask whether fear response behaviour in urban birds is associated with the differing constraints of the breeding and non-breeding seasons. To examine this, we measured the escape behaviour of an urban, resident population of Dark-eyed Juncos (Junco hyemalis). We quantified escape behaviour from 2023-2025 by measuring the flight initiation distance (FID) of individually color-banded juncos and compared junco FID throughout the year. We found that junco FID was significantly reduced during the breeding season. Other significant predictors of FID in our model were observer starting distance and group size. Our results suggest that while seasonality plays a major role in determining urban species’ behavioural responses, other factors make this relationship more complex. During the breeding season, the perceived threat of predation may be diminished for juncos focusing on parental care, resulting in lower fear responses. Meanwhile, differing social dynamics during the non-breeding season such as flocking behaviour may cause complex effects on urban juncos’ fear response. Our study provides evidence that urban wildlife exhibit a flexible fear response that is seasonally dependent and related to breeding behaviour. These results help us more accurately understand the priorities of animals in urban environments, as well as how they continue to persist under changing conditions.
Dataset DOI: 10.5061/dryad.905qftv10
Description of the data and file structure
Files and variables
File: Seasonal_Fear_Code_Dryad.Rmd
Description: The markdown file shows the code with some explanations of the processes behind the code and statistical analysis.
File: Seasonal_Fear_Data_Jan29.csv
Description: The data used in the analysis.
Variables
Sex = sex of bird (male, female, or unknown if assayed during first summer of life).
USGS = the unique United States Geological Survey bird band number that was affixed to each bird's leg.
Age = age of bird. Hatch year (HY, first year of life), second year (SY, second year of life), after second year (ASY, after second year of life). Juncos are visually distinguishable until after their second year of life.
Observer = There were two observers conducting FID assays. Observer 1, and observer 2.
G.Date = The Gregorian date of the FID assay.
Time = The time of day that the assay was taken.
G. Sunrise = The time of sunrise in the Gregorian time system (ex. 8:00am).
J.Sunrise = The time of sunrise in the Julian time system (ex. 428 minutes since 0 midnight).
Minutes = The time that the FID assay was taken in the Julian time system.
Min.Sunrise = the difference in minutes between the time the FID was taken and the time that sunrise occurred that day.
Long.Date = The days passed since the 1st day of 2023, without resetting in January of 2024 (ex. Feb 11 2024 was the 408th day after January 1 2023.
Stacked.Date = The day of the year that the FID was taken (October 28th is the 301st day of the year).
J.Date = The Julian date on which the FID was taken (ex. 2023301 = the 301st day of 2023).
Year = The year that the FID was taken. The data was taken between 2023 and 2025.
Season = The season during which the FID was taken. There are two seasons, the non-breeding season and the breeding season. The breeding season is between March and July and the non-breeding season is between August-February.
Precise.Season = The season including the year (ex. Non-breeding23 took place during the 2023 non-breeding season). January 2024 counts as the 2023 non-breeding season because it is a continuation of the 2023 non-breeding season).
Month = the month during which the FID assay was taken.
Path.Dist = The distance from a human path in meters from where the bird was when the FID was taken.
Height = the height off the ground where the bird was when the FID was taken (meters).
People.Present = the number of other people present within 20 meters.
Group.Size = the total number of juncos present within 5 meters.
Group = Whether or not the junco was in a group or not (Flock or not).
Conspecifics = the number of other juncos within 5 meters.
Other.spp.present = whether or not there were other bird species present within 5 meters, and how many species if yes.
Other.Species = a note on how many birds of other species were present within 5 meters, and how many there were.
Precise.Breeding.Status = Whether or not we were aware of the bird actively having eggs, nestlings, or fledglings at the time of FID assay.
Breeding.Status = Whether or not the bird had bred at all that season yet or not. Reset each year for repeat birds.
Euclidean.FID = the flight initiation distance (meters), accounting for the height of the bird.
FID = the flight initiation distance (meters), which is the distance between the observer and the bird at the moment the bird fled from the observer.
Starting.Distance = the distance (meters) that the observer was when they first started approaching the bird at 0.5 m/s for the FID assay.
Wind = the wind speed (kilometers per hour) at the time of FID assay.
Temperature = the temperature (Celsius) at the time of the FID assay.
x = the x coordinates of the location of FID assay.
y = the y coordinates of the location of FID assay.
Code/software
RStudio version: Version 2024.12.0+467
Packages and versions:
rsample (ver. 1.2.1)
dplyr (ver. 1.1.4)
rpart (ver. 4.1.24)
rpart.plot (ver. 3.1.2)
ipred (0.9-15)
caret (ver. 7.0-1)
ggplot2 (ver. 3.5.1)
ggsignif (ver. 0.6.4)
ggforce (ver. 0.4.2)
Metrics (ver. 0.1.4)
cAIC4 (ver. 1.0)
lme4 (1.1-36)
lmtest (0.9-40)
tidyverse (ver. 2.0.0)
MuMIn (ver. 1.48.11)
car (ver. 3.1-3)
sjPlot (ver. 2.9.0)
multcomp (ver. 1.4-28)
Access information
Instructions to run files:
ensure that the RStudio working directory is a folder on your computer that has both the code file and the data file. Then you can use+ Cmd + Shift + Enter to run all the code in the Rmd.
