Adult survival is the most important demographic parameter influencing population dynamics for many bird taxa. Thus, understanding how survival probabilities and causes of mortality vary throughout the annual cycle is critical for developing informed and effective management strategies. In this study, we used radio-telemetry data to evaluate the effects of biotic (e.g., sex, peak [September–April] vs. off-peak [May–August] nesting seasons) and abiotic factors (e.g., rainfall, year, bi-monthly interval) on adult survival, estimate annual survival probabilities, and identify primary sources of mortality for Hawaiian Ducks (Anas wyvilliana), an endangered, non-migratory dabbling duck, on the island of Kaua‘i, Hawai‘i, USA over 2013 and 2014. Additionally, we used contemporaneous Hawaiian Duck carcass recovery and surveillance data to examine temporal and climatic associations with avian botulism outbreaks. Our results suggested bi-monthly survival decreased with total rainfall during the preceding two-month interval. Survival did not vary with sex, between peak and off-peak nesting seasons, or between the two years of this study. Annual survival probabilities (62–80%) were relatively low compared to the closely related Laysan Duck (Anas laysanensis) on Laysan Island. Primary causes of mortality included avian botulism and presumed predation by cats (Felis catus). The botulism surveillance dataset revealed support for the effect of rainfall on the number of sick and dead birds recovered (n = 216), with generally a greater number of recoveries during months with middle-range total rainfall during the concurrent and preceding months. Our study provides critical baseline demographic data for population monitoring and highlights the importance of managing botulism risk and non-native mammalian predators for the recovery of the endangered Hawaiian Duck.

We attached 18-g radio-transmitters (Model AI-2M, Holohil Systems, Carp, Ontario, Canada) to adult Hawaiian Ducks (n_female = 50, n_male = 67) at Hanalei National Wildlife Refuge on the north shore of Kaua‘i, Hawai‘i, USA. We monitored the location and status (i.e., alive, dead) of tagged birds from December 2012 through December 2014. When carcasses or remains were recovered, we classified the suspected cause of mortality as predation by mammal or bird, avian botulism, other pathogen, anthropogenic, or unknown using a combination of evidence at the mortality site, external examination of the carcass, necropsy, and testing for botulinum toxin type C. 

During our study, USFWS personnel conducted botulism surveillance at least once weekly and up to three times daily to recover sick birds and carcasses at Hanalei NWR. During botulism outbreaks, observers walked transects within affected impoundments and lo‘i. Sick birds and carcasses were also recovered incidentally while performing other tasks on the refuge.

We obtained rainfall and temperature data from USGS climate station USC00518165 at Princeville Ranch, ~1 km north of the refuge (National Climate Data Center, https://www.ncdc.noaa.gov). Missing data (4% for rainfall, 8% for temperature) were supplemented with data from nearby climate stations (USC00513386, US1HIKI0007, USC00514561).

Please review the associated article for detailed descriptions of response variables, covariates, and analyses. We provided file and data descriptions in a README.txt file.

Known-fate Survival Analysis

Required software: Program MARK (available at: http://www.phidot.org/software/mark/)

Required files: inp and csv files

Cause-specific Mortality Estimates

Required software: The R Project for Statistical Computing (available at: https://www.r-project.org/)

Required files: R and csv files

Botulism Analysis

Required software: The R Project for Statistical Computing (available at: https://www.r-project.org/)

Required files: R and csv files