Invasion dynamics of the European Collared-Dove are explained by combined effects of habitat and climate
Data files
Sep 19, 2023 version files 533.85 KB
-
DRYADDataBundle.zip
529.44 KB
-
README.md
4.41 KB
Abstract
Global biodiversity is increasingly threatened by the spread of invasive species. Understanding the mechanisms influencing the initial colonization and persistence of invaders is therefore needed if conservation actions are to prevent new invasions or strive to slow their spread. The Eurasian collared-dove (Streptopelia decaocto, EUCO) is one of the most successful avian invasive species in North America; however, to our knowledge, no study has simultaneously examined the role that climate-matching, human activity, directional propagation, and local density have in this invasion process. Our research expands upon a cellular-automata-based hierarchical model developed to assess directional invasion dynamics to further quantify the impacts of climate, elevation, and land cover type on the spread of EUCO in North America. Our results suggest that EUCO’s dispersal patterns can largely be explained by the effects of habitat, climate, and environmental conditions at different stages of the invasion process rather than some innate preferred north-westerly spread. Specifically, EUCO initially colonized warm and wet grassland habitats and tended to persist in urban areas. We also found that while EUCO were more likely to spread to the northeast of existing habitats, directional preference did not drive persistence and recolonization events. These findings highlight the importance of incorporating both neighbourhood effects and environmental factors in the modelling of range-expanding species, adding to the toolset available to researchers to model invasive species spread. Further, our research demonstrates that historical records of invasive species occurrences from citizen science projects can provide the data resources needed to disentangle the characteristics driving species invasion and enable predictions that are of critical importance to resource managers.
README: Data from: Invasion dynamics of the European Collared-Dove are explained by combined effects of habitat and climate
This dataset comprises Eurasian collared dove detection records from Project FeederWatch, enriched with geographic and environmental descriptors of each location. Our research expands upon a cellular-automata-based hierarchical model developed to assess directional invasion dynamics to quantifiably demonstrate how climate, elevation, and land cover impact the dove's spread.
Description of the data and file structure
Merged_allCells_obsCovLatlon_1998-2021_US&CA.csv
- OBJECTID (int): A unique identifier representing the specific grid cell within the Christmas Bird Count (CBC) structure, enabling spatial referencing.
- year (int): The year in which the observation was recorded, providing temporal context for the dove's presence.
- eucdov (int): The number of survys with Eurasian collared doves observations records within the grid cell during the given year, forming the core of the dataset.
- noEucdov (int): The total number of bird count surveys conducted within the grid cell during the specified year, serving as a measure of survey effort.
- Elev_Mean (float): The elevation of the grid cell's location, aiding in understanding altitudinal preferences of the species.
- Temp_Mean (float): Average temperature recorded within the grid cell, contributing to the analysis of climatic influences on species distribution.
- Prec_Mean (float): Total precipitation measured within the grid cell, an important climatic factor impacting the dove's habitat.
- forest_per (float): Percentage of forest cover within the grid cell, representing a key land cover type affecting the species' habitat.
- crop_per (float): Percentage of crop-covered area within the grid cell, offering insights into anthropogenic landscape impacts.
- shrub_per (float): Percentage of shrub cover within the grid cell, contributing to the assessment of diverse habitat preferences.
- urban_per (float): Percentage of urbanized land within the grid cell, highlighting potential interactions with urban environments.
- water_per(float): Percentage of water body within the grid cell, supporting the removal and evaluation of cells covered mostly with water.
- other_per (float): Percentage of other land cover types within the grid cell, providing a comprehensive view of habitat diversity.
- centroid_long (float): The longitude coordinate of the grid cell's center, complementing the latitude for accurate spatial referencing.
- centroid_lat (float): The latitude coordinate of the grid cell's center, aiding in precise geographic localization.
Sharing/Access information
Data was derived from the following sources:
- Project FeederWatch
- ScienceBase North America Climate data
- ScienceBase North America Elevation data
- North American Land Change Monitoring System (NALCMS) land cover map 2015
Code/Software
NIMBLEcode_fullModel_forPredict.txt
The main model in study, written in the BUGS language, which can be invoked and driven from R using the NIMBLE package.
nimbleDriver_FromRDS.R
The R code used to drive the models in the study, reads jags_data_file.RDS
as input which can be generated with GenerateJAGSData.R
from Merged_allCells_obsCovLatlon_1998-2021_US&CA.csv
and N1ForAll.csv
.
Package versions:
- nimble_0.13.1
GenerateJAGSData.R
The R code used to filter and reformat the observation data into a list that is readable by the R driver code that runs the models.
Package versions:
- tidyverse_2.0.0
N1ForAll.csv
A table listing the cell neighbourhood relationships. Each row represents a cell, with its row number (first column) indicating its OBJECTID. Each row has 8 columns, indicating the OBJECTIDs of its 8 direct neighbouring cells. An NA entry in this table indicates that a cell has no neighbour in that specific direction. This occurs when a cell lies along the boundary of the study area or borders a body of water.
Methods
Dataset Collection:
The primary dataset in this study consists of detection records of Eurasian collared doves (Streptopelia decaocto) spanning from 1988 to 2021 in North America. These records were sourced from the Project FeederWatch database. Accompanying these records are the total number of bird count surveys conducted over this period, enriched with geographic and environmental descriptors of each location.
Processing and Gridding:
Temporal Grouping: The detection records were grouped into annual recordings.
Spatial Gridding: Detection records were aligned using the grid cell structure adopted by the Christmas Bird Count (CBC).
Filtering: We filtered out grid cells that had no surveys, and cells that were predominantly water-covered (> 50% of their area).
Environmental Descriptors:
Climate Data: Average temperature and precipitation data were integrated from ScienceBase North America Climate data (1950–2000).
Elevation: Elevation data was sourced from ScienceBase North America Elevation data.
Land Cover: Land cover descriptors, which included forest, shrub, crop, urban, and other land cover types, were integrated using the North American Land Change Monitoring System (NALCMS) land cover map from 2015.