Shifts in avian migration phenologies do not compensate for changes to conditions en route in spring and fall

Abstract

Several factors are known to affect bird migration timing, but no study has simultaneously compared the effects of temperature, land surface phenology, vegetation greenness, and relative humidity in both spring and fall. In addition, it is unclear whether long-term shifts in migration phenologies have kept pace with changing climates. For example, if migration shifts earlier in the spring, temperatures on migration dates may remain stable over time despite spring warming trends. If the phenologies of birds, plants, and insects shift asynchronously in response to changing climates, then birds may encounter reduced resource availability during migration. We estimated spring and fall 10%, 50%, and 90% cumulative migratory passage dates at 53 weather surveillance radar stations across the US Central Flyway. We determined which conditions (temperature, timing of green-up and dormancy, relative humidity, and enhanced vegetation index [EVI]) explained annual variation in migration phenologies. We also described decadal trends in environmental conditions and whether shifts in migration phenologies were sufficient to compensate for these changes. Annual changes to spring migration phenologies were best explained by anomalies in temperature, with earlier passage in warmer years. Fall migration occurred later on warmer, more humid years with higher EVI and later dormancy. Long-term adjustments in bird migration phenologies did not mitigate their exposure to changing environmental conditions. Although passage dates for all spring migration quantiles advanced significantly (~0.6 days/decade), temperatures on spring 10% passage dates increased, while 50% and 90% passage occurred closer to green-up. In the fall, temperatures increased on 50% and 90% passage dates. By contrast, the advancement of 10% passage (~1 day/decade) prevented early migrants from experiencing the cooling late-summer temperature trend. Warmer temperatures in mid to late fall may lead to earlier fruiting phenology and asynchronies with migratory passage, which occurred later in warmer years. Changes in temperature and land surface phenophases experienced by migrants suggest that resource availability during migration has changed and that adjustments to migration phenologies have not compensated for the effects of changing climates.

Keywords: aeroecology; aves; birds; climate change; green‐up; land surface phenology; migration; phenology; radar ornithology.

FIGURE 1

Study area map, definition of passage periods, and hypothetical scenarios for how changes to migration phenologies could compensate for environmental changes experienced by migrants. Panel (a) shows the study areas and radar zones. Panels (b/c) show two stations' passage periods (dashed lines), defined by the earliest and latest day‐of‐year that each passage quantile occurred at each station across all years. Points represent passage dates. Solid lines and shaded areas show trends in passage dates and 95% credible intervals (CIs). Panels (d–g) show four scenarios for how shifting migration phenologies could prevent the assemblage of migrants from experiencing changes in environmental conditions that occur during passage periods.

FIGURE 2

Effects of interannual anomalies in environmental conditions on annual passage dates. Interannual anomalies were defined as the observed values of the environmental conditions minus their expected values based on long‐term trends. Panel (a) shows standardized effect sizes for the interannual anomalies in each environmental condition/season. Shaded areas show the posterior probability distributions, and error bars show the 95% credible intervals (CIs). Panel (b) shows the passage dates versus interannual temperature anomalies. The lines and shaded areas show the expected values and 95% CIs for passage dates at 37° N and 96° W. EVI, enhanced vegetation index.

FIGURE 3

Decadal trends in environmental conditions on passage dates (solid lines) and during passage periods (dashed lines). Shaded areas represent 95% credible intervals (CIs) for conditions on passage dates (darker shading) and during passage periods (lighter shading). If migration phenologies shifted to compensate for changes in environmental conditions (Figure 1d/e), we expected conditions on passage dates (solid lines) to remain unchanged or have a shallower slope than overall trends in conditions during the passage periods (dashed lines). Panels (d/h) include trend lines for green‐up and dormancy dates. Predictions are shown for 37° N and 96° W. EVI, enhanced vegetation index.