Data from: Increased mortality rates caused by highly pathogenic avian influenza virus in a migratory raptor
Data files
Jul 17, 2025 version files 44.08 MB
-
AnnualMortality.csv
14.57 KB
-
AnnualMortalityComparison.r
4.50 KB
-
Buteo_lagopus_Conway-Paprocki_NA.csv
3.65 MB
-
Figure_2_code.r
4.41 KB
-
Figure_3_sex_analysis.r
3.59 KB
-
Kidd_et_al._Rough-legged_Hawk_Movements_in_North_America.csv
40.39 MB
-
mortality-summary-RLHA.csv
3.72 KB
-
README.md
4.50 KB
-
sexing_Buteo_HPAI.csv
2.80 KB
Abstract
Highly pathogenic avian influenza virus (HPAIV) has caused extensive mortalities in wild birds, with a disproportionate impact on raptors since 2021. The population-level impact of HPAIV can be informed by telemetry studies that track large samples of initially healthy, wild birds. We leveraged movement data from 71 rough-legged hawks (Buteo lagopus) across all major North American migratory bird flyways concurrent with the 2022–2023 HPAIV outbreak and identified a total of 29 mortalities, of which 11 were confirmed, and an additional ~9 were estimated to have been caused by HPAIV. We estimated a 28% HPAIV cause-specific mortality rate among rough-legged hawks during a single year concurrent with the HPAIV outbreak in North America. Additionally, the overall annual mortality rate during the HPAIV outbreak (47%) was significantly higher than baseline annual mortality rates (3–17%), suggesting that HPAIV-caused deaths were additive above baseline mortality levels. HPAIV mortalities were concentrated within the Central and Atlantic flyways during pre-breeding migration and peaked in April 2022 when large-scale HPAIV mortalities were reported in other wild birds throughout North America. HPAIV exposure was most likely caused by scavenging or preying on infected waterfowl, as rough-legged hawks are known to opportunistically scavenge during the nonbreeding season. We utilized movement data to identify a continental-scale HPAIV cause-specific mortality event in rough-legged hawks that has the potential to exacerbate ongoing population declines. Our study highlights the usefulness of monitoring movement data to pinpoint sources of mortality that can help better understand the drivers of population change, even if studies are focused on other research questions.
Dataset DOI: 10.5061/dryad.n2z34tn92
Description of the data and file structure
We leveraged movement data from GPS-tracked rough-legged hawks Buteo lagopus that coincided with the HPAIV panzootic in North America to determine its effect on annual mortality. All missing and unavailable data represented as NA
Files and variables
File: AnnualMortality.csv
Description: spreadsheet used to analyze the HPAIV effect on annual mortality.
Variables
- index: index number
- tagid: transmitter ID (unique to an individual hawk)
- year: 12-month study period. 2020 = 1-Mar-2020 to 28-Feb-2021, etc…
- date_begin: date within the 12-month study period that tracking began
- date_end: date within the 12-month study period that tracking ended
- duration: tracking duration (number of days)
- fate: individual fate during the 12-month study period. See the manuscript methods for details.
- COD: cause of death. See the manuscript methods for details.
- tagtype: transmitter type - PTT or GSM.
- Sex: individual sex
- day12mo: day within the 12-month study period that tracking began. 1-Mar = day 1
File: AnnualMortalityComparison.r
Description: R script for comparison of annual mortality estimates.
File: Figure_2_code.r
Description: R script for Figure 2 code.
File: Figure_3_sex_analysis.r
Description: R script for Figure 3 code and HPAIV sex/age analysis.
File: Buteo_lagopus_Conway-Paprocki_NA.csv
Description: Buteo lagopus movement data used in Figure 2.
Variables
- event-id: row identifier.
- Timestamp: date and timestamp of bird GPS location.
- Location-long: longitude associated with the GPS timestamp.
- Location-lat: latitude associated with the GPS timestamp.
- Sensor-type: location type - GPS or Argos.
- individual-taxon-canonical-name: species name
- individual-local-identifier: transmitter ID
- study-name: movebank study name
File: mortality-summary-RLHA.csv
Description: data used for Figure 2 and Figure 3.
Variables
- tagid: transmitter ID
- fate: individual fate during 12-mo study period. See the manuscript methods for details.
- COD: cause of death. See the manuscript methods for details.
- Plotting: COD reformatted for graphing/plotting.
- last_latitude: last known latitude.
- last_longitude: last known longitude.
- sex: F = female; M = male
- date_end: end date of tracking period for tagid.
File: sexing_Buteo_HPAI.csv
Description: data used for age or sex effects on cause-specific mortality (HPAIV positive vs. all others)
Variables
- Species: species 4-letter code
- tagid: transmitter ID
- mass: body mass (grams) at capture.
- Wing: wing chord (mm) at capture.
- ageCapture: age at capture. 1C = first cycle; 2C = second cycle; A = older than second cycle.
- fate: individual fate during the 12-month study period. See the manuscript methods for details.
- COD: cause of death. See the manuscript methods for details.
- Age: age at end of study period. J = juvenile/first cycle; A = adult/older than first cycle
- sex: F = female; M = male.
File: Kidd_et_al._Rough-legged_Hawk_Movements_in_North_America.csv
Description: Buteo lagopus movement data used in Figure 2.
Variables
- event-id: row identifier.
- Timestamp: date and timestamp of bird GPS location.
- Location-long: longitude associated with the GPS timestamp.
- Location-lat: latitude associated with the GPS timestamp.
- Sensor-type: location type - GPS or Argos.
- individual-taxon-canonical-name: species name
- individual-local-identifier: transmitter ID
- study-name: movebank study name
Code/software
We used logistic regression via the ‘glm’ function from the ‘stats’ R package (R Core Team, 2025) to predict HPAIV fate explained by sex and age class.
We used logistic regression via the ‘glm’ function from the ‘stats’ R package (R Core Team, 2025) to predict the fate (1 = ‘likely dead’ or ‘confirmed dead’; 0 = ‘alive’ or ‘unknown’) of tracked birds across four, 12-mo periods. We used Tukey pairwise comparisons from the ‘emmeans’ package (Lenth, 2023) to determine pairwise differences in mortality rates among years.
Access information
Other publicly accessible locations of the data:
- N/A
Data was derived from the following sources:
- N/A