Data and code from: Urbanisation and human activities influence the co-occurrence of red squirrels Sciurus vulgaris and meso-carnivores in Berlin, Germany
Data files
Mar 12, 2026 version files 16.33 MB
-
README.md
4.37 KB
-
squirrels_msom_vergara.7z
16.32 MB
Abstract
Urban expansion is a major driver of habitat fragmentation, shrinking wildlife habitat, and restricting wildlife movements and activity patterns. In this novel environment, species must adapt to the new composition of wildlife communities. For example, red squirrels (Sciurus vulgaris) are commonly found in urban environments, while their potential predators have also managed to accommodate and even thrive in cities. Consequently, studying species interactions in an urban landscape provides crucial insights into our understanding of species’ plasticity and behavioural adjustments to novel situations. This study aims to understand the potentially novel predator–prey interactions in private gardens of Berlin, specifically between red squirrels and red foxes Vulpes vulpes, raccoons (Procyon lotor), martens (Martes sp.), and cats (Felis catus). We hypothesized that squirrels’ occurrence would be influenced by human activity, garden characteristics, and proximity to the city centre, and that these factors would also affect co-occurrence with meso-carnivores. We used a raster of 150 camera traps per season, in total 575, located in private gardens of citizens of Berlin, which detected the species during four sampling seasons between 2019 and 2020. We first applied single occupancy models for each species separately to understand the variation in the detection and occupancy of each species at given environmental variables. We then assessed paired co-occurrence of squirrels with each meso-carnivore by applying a multi-species occupancy model framework. We found that squirrels were more likely to occur during autumn than in spring and at a higher percentage of local tree cover, which has been confirmed by previous studies. Additionally, results suggested that squirrels are more likely to occur when cats were absent, suggesting an avoidance of the latter. However, squirrels did not display clear patterns of negative or positive co-occurrence with other urban wild meso-carnivores. These results demonstrate the complexity of species interactions in urban settings.
Dataset DOI: 10.5061/dryad.hmgqnk9x8
Description of the data and file structure
Overview
This dataset contains camera trap records from a study investigating occupancy and species co-occurrence between red squirrels (Sciurus vulgaris) and urban mesocarnivores in Berlin, Germany.
Data Collection
- Location: Berlin, Germany
- Habitat type: Private urban gardens
- Sampling periods:
- Spring 2019
- Autumn 2019
- Spring 2020
- Autumn 2020
- Equipment: Motion-triggered camera traps
- Number of sites: ~600 locations across four sampling periods
- Target species:
- Red squirrel (Sciurus vulgaris)
- Domestic cat (Felis catus)
- Red fox (Vulpes vulpes)
- Marten (Martes martes and Martes foina)
- Raccoon
Files and variables
File: squirrels_msom_vergara.7z
Description:
R project: squirrels_msom_vergara_j_s.Rproject
Folders:
- data-raw: raw data to run the code.
- data_all.csv: Camera trap detection records from urban gardens, including species identification, date and time of detection, and site characteristics.
- env_var.csv: Local environmental variables
- stacked_raster_values_CT_all_seasons.csv: City scale environmental variables
- Output: folder where output will be stored.
- plots: folder where plots resulting from the code will be stored.
- R: script from the project.
Variables [unit/categories] per tabular data:
- data_all.csv:
- id_season: survey season [1, 2, 3, 4]
- determination: species identificated [biber, cat, dachs, eichhörnchen, feldhasen, igel, marder, maus, ratte, reh, rotfuchs, waschbär, wildkaninchen, wildschwein]
- date: begining date of the record [dd/mm/yyy]
- time: begining time of the record [hh:mm:ss]
- crdate: end date of the record [dd/mm/yyy]
- crtime: end time of the record [hh:mm:ss]
- User_uid: identification number of the garden/user
- dog_presence: presence of dogs in gardens [Dog present - Dog in house at night -No dog]
- garden_size: Garden size [m^2^]
- local_tree_cover: tree cover in gardens [%]
- compost: compost presence in gardens [Non - Open - Close]
- fence_height: fence height in gardens [cm]
- env_var.csv:
- covid: lockdown due to SARS-CoV2 [0 (no lockdown) - 1 (during lockdown)]
- season: season [spring - autumn]
- stacked_raster_value_CT_all_seasons.csv:
- DistanceCityBorder: distance to city boreder [m]
- Focal_pop_100m: human population density [inhabit/ha]
- distance_water: distance to water [m]
- impervious_surface: impervious surface [%]
- noise: noise level [dBa]
- tree_cover: tree cover [%]
Abreviations [unit/categories] in the code:
- IDS: survey season [1, 2, 3, 4]
- lcltree: Local tree cover [%]
- cityBorder: Distance to city boreder [m]
- dog: Dog presence [Dog present - Dog in house at night -No dog]
- cvd: Lockdown due to SARS-CoV2 [0 (no lockdown) - 1 (during lockdown)]
- garden: Garden size [m^2^]
- ssn: Season [spring - autumn]
- water: Distance to water [m]
- pop: Human population density [inhabit/ha]
- noise: noise level [dBa]
- imp: Impervious surface [%]
- fence: Fence height [cm]
- NA: Not available
Missing data: Some cells contain missing values because the citizen science contributor did not complete all required fields or because the information provided was incomplete or incorrectly recorded. Missing values are coded as "NA". This format is retained because the analysis scripts used automatically interpret "NA" as missing data.
Translation
Since some variables and categories are in German, the following are the translations into English:
- biber: beaer
- dachs: badger
- eichhörnchen: squirrel
- feldhasen: hare
- igel: hedgehog
- marder: marten
- maus: mouse
- ratte: rat
- reh: deer
- rotfuchs: red fox
- waschbär: raccoon
- wildkaninchen: wild rabbit
- wildschwein: wild boar
- kein tier: no animal
- immer freigang: always free
- nachts hause: house at night
- offen: open
- geschlossen: closed
- keiner: non
Code/software
RStudio Team. (2020). RStudio: Integrated Development for R (Version 2022.12.0+353) [Computer software]. RStudio, PBC. http://www.rstudio.com/
