Résumé
Non-additive effects from multiple interacting stressors can have
unpredictable outcomes on wildlife. Stressors that initially have
negligible impacts may become significant if they act in synergy with
novel stressors. Wildlife markers can be a source of physiological stress
for animals and are ubiquitous in ecological studies. Their potential
impacts on vital rates may vary over time, particularly when changing
environments impose new stressors. In this study, we evaluated the
temporal changes in the combined impact of two stressors, one constant
(collar-marking) and another one variable over time (hunting intensity),
in greater snow geese (Anser caerulescens atlantica). Over a 30-year
period (1990-2019), hunting regulations were liberalized twice, in 1999
and 2009, with the instauration of special spring and winter hunting
seasons, respectively. We evaluated the effect of collars on goose
survival through this period of changing hunting regulations. We compared
annual survival of >20,000 adult females marked with and without
neck collars using multievent capture-recapture models, and partitioned
hunting from non-hunting mortality. Survival of geese marked with or
without collars was similar in 1990-1998, before hunting regulations were
liberalized (average survival[95% C.I.]: 0.87[0.84, 0.89]). However,
absolute survival of collared geese was 0.05[0.03, 0.07] lower than that
of non-collared geese between 1999 and 2009, and 0.12[0.09, 0.15] lower
after hunting regulations were liberalized further in 2009. Hunting and
non-hunting mortality were both higher in collared birds compared to those
without collars. The interaction between the effects of collars and
hunting was synergistic because collars affected survival only after the
hunting pressure increased significantly. These cumulated stresses
probably reduced goose body condition sufficiently to increase their
vulnerability to multiple sources of mortality. Synthesis and
applications: Researchers relying on long-term marking programs should
evaluate the effect of markers periodically rather than solely in the
beginning, as interactions with changing environmental conditions may
eventually affect conclusions of studies based on marked animals. Here, we
provide a rare demonstration in a natural setting that a combination of
stressors can push animals beyond a threshold where vital rates are
affected, even when one stressor applied alone initially had no detectable
impact.