Substantial global data show shifting phenologies across many organisms in response to climate change. For birds, migration arrival dates in breeding regions have been shifting earlier and there is evidence both adaptation and plasticity influence these shifts. We hypothesize that flight efficiency, as measured by morphology of the flight apparatus, may be associated with larger shifts, as better fliers may be able to better respond to changing conditions during migration. The use of large-scale citizen science data, such as eBird, presents a powerful tool to investigate arrival date shift, as data are analogous to amateur birding checklists historically used in migration tracking. As the usage of these programs has increased, a methodology addressing variable sampling effort needs to be applied, as looking at early arrivals on data with increasing sampling effort over time may bias arrival date trends. We applied a logistic model to 12 years of eBird data for 35 common passerines in the Greater Toronto Area to estimate mean arrival date (MAD). We then used linear regression to analyze changes in MAD with time, migratory distance, and flight ability.
Results/Conclusions:
We find that of the 35 species analysed, 22 showed earlier arrivals (β = -0.03 to -0.80) and 13 showed later arrivals (β = 0.01 to 0.32) during the study period. Overall, birds arrived approximately 3.5 days earlier in 2019 than 2008 (β = -0.29, p-value = 0.07, r2 = 0.29), a rate of change similar to those reported globally. Short distance migrants had a stronger shift (β = -0.44, p-value = 0.12, r2 = 0.22) than long distance migrants (β = -0.21, r2 = 0.20, p=0.15). A weak relationship existed between flight efficiency proxies and MAD, suggesting that plasticity in migration timing may be affected by morphology. In an increasingly warming world, tracking such phenological changes will be instrumental in understanding which species are most at risk, particularly as phenological mismatches are associated with population declines. Large-scale citizen science data presents the ability to model and track these changes across greater geographic ranges and with greater sampling effort than past data collection schemes.