Data and code from: Forest bird body condition and fruit consumption related to Spotted wing Drosophila abundance and fruit parasitism
Data files
Jun 09, 2026 version files 589.25 KB
-
allbirdcaptures.csv
160.16 KB
-
cutdata.csv
38.24 KB
-
forestdata.csv
16.97 KB
-
guildbirdcaptures.csv
17.02 KB
-
README.docx
19.52 KB
-
README.md
6.36 KB
-
round1-Nmixcode.R
89.94 KB
-
SWD_Metadata.xlsx
13.80 KB
-
SWDsummary.R
227.24 KB
Abstract
Migratory bird species are under heavy energetic demands following the breeding season. Birds must first complete a pre-migratory molt and build fat deposits during the post-breeding season for migration. Many birds utilize early successional habitat, consuming fruit resources for energy, and using cover provided by vegetation. Spotted wing Drosophila (Drosophila suzukii, SWD) is a frugivorous non-native invasive insect pest in the United States that parasitizes the ripening fruits of many wild fruiting plants which could affect birds reliant on fruits during a critical time of year. We evaluated what factors, especially those related to SWD abundance and parasitism, related to body condition of several bird species and probability of consumption of fruits by frugivorous birds in regenerating harvests invaded by SWD during the post-breeding season and fall migration. We used mist nets to catch birds in regenerating harvests at Allegheny National Forest from July to October in 2019 and 2020 and measured multiple indices of body condition (subcutaneous fat, molt progression, and scaled mass index [SMI]) for several species. We also collected fecal samples from several frugivorous species for evidence of fruit consumption. We assessed fruit resources and vegetation structure at each mist net using transect surveys. We trapped SWD at each mist net for the duration of mist netting sessions and collected fruit samples to determine the extent of fruit parasitism by SWD at each mist net. Multiple condition indices for several species were correlated with relative abundance of SWD and parasitism of fruits, though the directions of relationships varied and were species-specific. Probability of Allegheny blackberry (Rubus allegheniensis) fruit consumption by frugivorous birds was positively correlated with parasitism of fruits at both the guild and species level. Most species improved condition indices as the season progressed, likely due to dietary flexibility. Future research should include how bird condition changes in relation to foraging microhabitat selection and SWD, selection or rejection of SWD-parasitized fruits by frugivores, use of nutritional metrics to assess quality of SWD and SWD-parasitized fruits, arthropod community responses to SWD, and the possible long-term consequences of changes to fruit consumption and seed dispersal.
Dataset DOI: 10.5061/dryad.zw3r228nw
For all spreadsheets, please reference the SWD_Metadata spreadsheet for definitions and details. For details about data collection protocols, please reference the methods section as part of the published manuscript.
Description of the data and file structure
Files and variables
File: SWD_Metadata.xlsx
Description:
metadata file containing definitions & details about variables in the other spreadsheets
Data Spreadsheets:
File: allbirdcaptures.csv
Description:
raw bird capture data. Regenerating young forest stands were sampled with 12-m mist nets for migrating songbirds during the fall of 2019 and 2020. Sites were in north-central Pennsylvania, USA. Each stand was sampled 3 times. The spreadsheet includes a stand identifier (site), the year, the point in each stand, and the species code including the age of the individual captured. The data are structured so that each random sampling point (represented by mist nets) within a young forest stand, receives a row of information. The sample unit is a mist net (random sample point) within a forest stand.
File: cutdata.csv
Description:
pertinent data related to sampling the fruiting vegetation in the young forest stands. These data are the explanatory (independent) variables that we used when modeling fruit and vegetation resources. The data are structured so that each random sampling point (represented by mist nets) within a young forest stand, receives a row of information. The sample unit is a mist net (random sample point) within a forest stand.
File: forestdata.csv
Description:
data pertinent to the cherry tree basal area and spotted winged drosophila sample information. The sample unit is a mist net (random sample point) within a forest stand.
File: guildbirdcaptures.csv
Description:
species are grouped by guild and age class and each row represents a sample unit, which is a mist net point in a forest stand.
File: SWDSummar.R
Description:
RCode for summarizing the excel spreadsheets and generating summary statistics.
File: round1-Nmixcode.R Rfile, see below.
Response variables:
Counts and Inclusion Rules
For each guild/age or species/age group, counts include:
- Newly banded birds
- Recaptures from previous years (first captured in 2019)
- Recaptures of birds only if they were captured at the same net on the second day
Important rule:
If a bird was caught in multiple nets on the same day, only the first capture was used.
Site Covariates:
Pbprop (Postbreeding proportion)
Proportion of total captures that met criteria for postbreeding.
Calculation
- For each guild/age or species/age group, counts are summed over each two-day sampling period
- Pbprop = number of postbreeder detections/captures ÷ total detections/captures
Reasoning
Site-level covariates can only have one value per primary sampling period, so values could not be assigned per secondary sampling period.
Postbreeding criteria
- HY birds
- AHY birds with:
- wrinkled/molting BP
- flight feather (ff) molt
- body molt score > 2 (25%)
Exceptions
- HY models do not include this covariate (all HYs are assumed postbreeding)
- Species richness models do not include Pbprop (species cannot be assigned breeder/postbreeder status)
Forest_type (categorical)
- 0 = oak-hickory
- 1 = mixed upland
- 2 = Allegheny hardwood
Fruit_abund
Total abundance of ripening + ripe Rubus allegheniensis fruits across both sampling transects at each mist net.
Fruit_prop
Proportion of total R. allegheniensis fruit that are ripening or ripe at each mist net.
- Calculated using pooled fruit abundance from both transects per net
SWD_abund
Abundance of spotted wing drosophila (SWD) at each mist net.
- Based on pooled trap counts from both traps per net
- Includes both male and female SWD
Infestation_rate
Number of adult SWD per gram of R. allegheniensis fruit.
- SWD counts and fruit mass are pooled across both transects per net
- Excludes larvae, pupae, pupal shells, and unknown life stages
Veg_height
Average vegetation height at each mist net.
- Mean of 6 measurements per net
- Each transect has measurements at 3 m, 6 m, and 9 m (2 transects × 3 points)
Basal_area
Basal area of all trees within 25 m of each mist net (m²).
- Calculated from DBH (cm) using:
BA = (π / (4 × 10000)) × DBH²
Cover_2.5m
Average vegetation cover (2–2.5 m height stratum) at each mist net.
- Mean of 12 measurements per net
- Each transect: 3 m, 6 m, 9 m on both sides (2 transects × 3 points × 2 sides)
Detection Covariates
Ordinal_date
Day of year (0–365)
Netday
Sampling day within the 2-day period:
- 0 = Day 1
- 1 = Day 2
Wind
Wind score at net opening (Beaufort scale: 0–3)
Temp
Temperature at net opening (°C)
Cloud
Cloud cover at net opening (%)
Precip
Precipitation score at net opening (0–5 categorical scale)
Humidity
Humidity at net opening (%), from local weather stations
Analyses
Global N-mixture models were run for:
Community-level models
- Frugivore HY, AHY, and total abundance
- Nonfrugivore HY, AHY, and total abundance
- Frugivore HY, AHY, and total species richness
- Nonfrugivore HY, AHY, and total species richness
Species-level models
- BTBW HY, AHY, and total abundance
- BTNW HY, AHY, and total abundance
- CEDW HY, AHY, and total abundance
- COYE HY, AHY, and total abundance
- CSWA HY, AHY, and total abundance
- EATO HY, AHY, and total abundance
- GRCA HY, AHY, and total abundance
- MAWA HY, AHY, and total abundance
- OVEN HY, AHY, and total abundance
- REVI HY, AHY, and total abundance
- SCTA HY, AHY, and total abundance
- SOSP HY, AHY, and total abundance
Model Distributions
For each response variable, three global models were fitted:
- Poisson
- Negative binomial
- Zero-inflated Poisson
Code/software
We used R Studio for analyses. The code for analyses and modeling is also included as a separate file named SWDsummary.R
