Samantha Bowser, Maggie MacPhersonPlease use the format "First name initials family name" as in "Marie S. Curie, Niels H. D. Bohr, Albert Einstein, John R. R. Tolkien, Donna T. Strickland"
<p>The acoustic adaptation hypothesis posits that animal sounds are influenced by the habitat properties that shape acoustic constraints (Ey and Fischer 2009, Morton 2015, Sueur and Farina 2015).Alarm calls are expected to signal important habitat and receiver-dependent information (Ripmeester et al. 2010, Sheldon et al. 2020), and we want to test whether Q. mexicanus alarm calls differ between populations and ecological contexts across the US as expected under the acoustic adaptation hypothesis (three US subspecies: Q. m. nelsoni, Q. m. monsoni, and Q. m. prospidicola; Figure 1). The alarm call vocalization in Q. mexicanus is known to vary in tone, range and pitch (Kok 1971). Alarm calls signal low intensity excitement (Kok 1971) and research in other species has shown that differences in the acoustic qualities of alarm calls reflect the urgency of threats tailored to the receiving audience (Carlson et al. 2020, Sheldon et al. 2020, McLachlan and Magrath 2020). However, due to the ecological importance of alarm calls in minimizing risk to group members, natural selection could promote stabilizing selection on alarm calls, resulting in homogenous alarm call structure across subspecies regardless of habitat and receiver. For this reason, we will also test whether Q. mexicanus songs differ between populations and ecological contexts across the US as natural selection likely promotes disruptive selection on song structure to facilitate subspecies recognition during mating season (Cruz-Yepez et al. 2020, Simpson et al. 2021). In this project we will enhance our understanding of the vocal repertoire of Q. mexicanus, by 1) recording and describing alarm calls and songs, 2) testing a null hypothesis that differing vocalizations will correlate with subspecies-specific soundscapes, and 3) test an alternative hypothesis that vocal signal characteristics correlate with range expansion. We will improve the description of vocalizations by recording vocalizations from each subspecies and analyzing the tone, range and pitch of vocalizations using spectrograms generated with Raven Lite 2.0 (Cornell Lab of Ornithology). Recording of alarm calls will take place during the non-breeding season, and of songs during the breeding season. We will only record alarm calls during the non-breeding period to avoid differences associated with reproduction. For our first objective, a phylogenetic principal component analysis (PPCA) will be conducted to identify correlations among measures of vocalization structure across subspecies while accounting for phylogenetic history. For our second objective, a phylogenetic generalized least squares analysis (PGLS) will be conducted to determine if subspecies vocalization characteristics are explained by social and habitat contexts within a phylogenetic context. To test whether vocalizations have functionally diverged and to help explain differences in range expansion, we will conduct a reciprocal playback experiment measuring responsiveness to recordings from within each subspecies compared to those from other subspecies. We will use the results of the PPCA and playback experiment to test whether vocal signal characteristics (both signal and response) are significant regional drivers of predicted distributions for Q. mexicanus in the US using an ensemble distribution model. If vocal signal skill is learned from context-dependent experiences unique to each subspecies (i.e., in line with the acoustic adaptation hypothesis), then individuals should share vocal characteristics with and respond to the signals of their own subspecies but not to signals of other subspecies. Tone, range, and pitch of vocalizations as well as low responsiveness will be a significant explanatory variable in all regional models (i.e., differences in vocal signals will distinguish subspecies distributions). However, if differences in regional models are due to variation in responsiveness according to subspecies, then skill in vocal communication could contribute to differences in range expansion among subspecies. Generalized linear models will be used to analyze the data of the playback experiment. All quality control processing and statistical analyses will be performed in the R environment (Araya‐Salas and Smith‐Vidaurre 2017, R Core Team 2020, Araya-Salas 2021). Data collection will stop once the minimum sample size is reached (n=88 per population, per season, calculated using a power analysis to detect small differences in signal strength). Quantifying differences in signal strength of vocalizations between Q. mexicanus subspecies will contribute to our understanding of vocal plasticity in heterogeneous landscapes. We will additionally test a series of alternative hypotheses using an ensemble distribution modelling approach to determine whether the distributions of each subspecies in the US are correlated with habitat variables (i.e., wetlands that do not freeze in winter, proximity of wetlands to disturbed habitat) and co-occurrence data with other blackbird species (i.e., occurrence data will be sourced from the citizen science platform, eBird).</p>
Biogeography, Biological invasions, Coexistence, Dispersal & Migration, Habitat selection, Landscape ecology
