Hunter harvest affects survival of Atlantic brant

Abstract

Adult survival is a key driver of population dynamics in long-lived species like geese. To implement effective and sustainable hunting regulations, managers need accurate estimates of survival and knowledge about the impacts of hunter harvest, particularly the extent to which this mortality is compensatory or additive to other sources. Atlantic brant (Branta bernicla hrota) are among the smallest of the North American goose populations and are vulnerable from a conservation perspective because of their limited breeding and wintering ranges, variable and low productivity, and reliance on coastal marine ecosystems. Additionally, the effects of hunter harvest on Atlantic brant survival are not fully known. We conducted both dead recovery only and joint live-dead mark-recapture analyses of Atlantic brant banded on Baffin Island and Southampton Island, Nunavut, Canada, during 2000–2018 to provide contemporary survival estimates and determine the extent to which hunter harvest was compensatory or additive. Survival probabilities of juveniles were lower and more variable ($\overline{x}$ = 0.54 ± 0.13 [SE]; range = 0.33–0.88) than adults ($\overline{x}$ = 0.84 ± 0.06; 0.74–0.95) and were influenced to a greater extent by non-harvest mortality. We found evidence of harvest additivity in adult Atlantic brant. Annual harvest of adult Atlantic brant explained 75% of the annual variability in adult survival probabilities, and the estimated process correlation between adult annual survival and recovery probabilities from a Brownie dead recovery model was negative (ρ = −0.34; SD = 0.32). Compared to other North American goose populations, Atlantic brant have lower harvest potential and less ability to compensate for hunter harvest. To ensure harvest remains sustainable, we suggest that managers should account for minimal harvest compensation in adult brant when selecting hunting regulations (i.e., season length and daily bag limits). Lastly, we encourage the use of joint live-dead models when possible because they provide greater insight into demographic processes and improved precision and accuracy on parameter estimates, particularly for juveniles.